Rename binary operation micro-kernels
PiperOrigin-RevId: 305834937
diff --git a/src/f32-vbinary/gen/vadd-avx-x16.c b/src/f32-vbinary/gen/vadd-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vadd-avx-x16.c
rename to src/f32-vbinary/gen/vadd-minmax-avx-x16.c
index 230e99f..ca894a0 100644
--- a/src/f32-vbinary/gen/vadd-avx-x16.c
+++ b/src/f32-vbinary/gen/vadd-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vadd_ukernel__avx_x16(
+void xnn_f32_vadd_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-avx-x8.c b/src/f32-vbinary/gen/vadd-minmax-avx-x8.c
similarity index 98%
rename from src/f32-vbinary/gen/vadd-avx-x8.c
rename to src/f32-vbinary/gen/vadd-minmax-avx-x8.c
index d89d07b..afb7989 100644
--- a/src/f32-vbinary/gen/vadd-avx-x8.c
+++ b/src/f32-vbinary/gen/vadd-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vadd_ukernel__avx_x8(
+void xnn_f32_vadd_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-avx512f-x16.c b/src/f32-vbinary/gen/vadd-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-avx512f-x16.c
rename to src/f32-vbinary/gen/vadd-minmax-avx512f-x16.c
index f4de3f0..30014fc 100644
--- a/src/f32-vbinary/gen/vadd-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vadd-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__avx512f_x16(
+void xnn_f32_vadd_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-avx512f-x32.c b/src/f32-vbinary/gen/vadd-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-avx512f-x32.c
rename to src/f32-vbinary/gen/vadd-minmax-avx512f-x32.c
index 3e80c7f..b7908c6 100644
--- a/src/f32-vbinary/gen/vadd-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vadd-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__avx512f_x32(
+void xnn_f32_vadd_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-neon-x4.c b/src/f32-vbinary/gen/vadd-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-neon-x4.c
rename to src/f32-vbinary/gen/vadd-minmax-neon-x4.c
index 259e053..6941dfe 100644
--- a/src/f32-vbinary/gen/vadd-neon-x4.c
+++ b/src/f32-vbinary/gen/vadd-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__neon_x4(
+void xnn_f32_vadd_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-neon-x8.c b/src/f32-vbinary/gen/vadd-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-neon-x8.c
rename to src/f32-vbinary/gen/vadd-minmax-neon-x8.c
index 1093b15..8faffed 100644
--- a/src/f32-vbinary/gen/vadd-neon-x8.c
+++ b/src/f32-vbinary/gen/vadd-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__neon_x8(
+void xnn_f32_vadd_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-psimd-x4.c b/src/f32-vbinary/gen/vadd-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-psimd-x4.c
rename to src/f32-vbinary/gen/vadd-minmax-psimd-x4.c
index 6fbfbf8..d72d81f 100644
--- a/src/f32-vbinary/gen/vadd-psimd-x4.c
+++ b/src/f32-vbinary/gen/vadd-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__psimd_x4(
+void xnn_f32_vadd_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-psimd-x8.c b/src/f32-vbinary/gen/vadd-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-psimd-x8.c
rename to src/f32-vbinary/gen/vadd-minmax-psimd-x8.c
index 9f70c26..05acd7f 100644
--- a/src/f32-vbinary/gen/vadd-psimd-x8.c
+++ b/src/f32-vbinary/gen/vadd-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__psimd_x8(
+void xnn_f32_vadd_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-scalar-x1.c b/src/f32-vbinary/gen/vadd-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vadd-scalar-x1.c
rename to src/f32-vbinary/gen/vadd-minmax-scalar-x1.c
index 4cab157..942b7b7 100644
--- a/src/f32-vbinary/gen/vadd-scalar-x1.c
+++ b/src/f32-vbinary/gen/vadd-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__scalar_x1(
+void xnn_f32_vadd_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-scalar-x2.c b/src/f32-vbinary/gen/vadd-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vadd-scalar-x2.c
rename to src/f32-vbinary/gen/vadd-minmax-scalar-x2.c
index 507e923..142a867 100644
--- a/src/f32-vbinary/gen/vadd-scalar-x2.c
+++ b/src/f32-vbinary/gen/vadd-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__scalar_x2(
+void xnn_f32_vadd_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-scalar-x4.c b/src/f32-vbinary/gen/vadd-minmax-scalar-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-scalar-x4.c
rename to src/f32-vbinary/gen/vadd-minmax-scalar-x4.c
index 6931e95..de7bad3 100644
--- a/src/f32-vbinary/gen/vadd-scalar-x4.c
+++ b/src/f32-vbinary/gen/vadd-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__scalar_x4(
+void xnn_f32_vadd_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-sse-x4.c b/src/f32-vbinary/gen/vadd-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-sse-x4.c
rename to src/f32-vbinary/gen/vadd-minmax-sse-x4.c
index a41028e..7b0e4e8 100644
--- a/src/f32-vbinary/gen/vadd-sse-x4.c
+++ b/src/f32-vbinary/gen/vadd-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__sse_x4(
+void xnn_f32_vadd_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-sse-x8.c b/src/f32-vbinary/gen/vadd-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-sse-x8.c
rename to src/f32-vbinary/gen/vadd-minmax-sse-x8.c
index 8161ed0..782f4d7 100644
--- a/src/f32-vbinary/gen/vadd-sse-x8.c
+++ b/src/f32-vbinary/gen/vadd-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__sse_x8(
+void xnn_f32_vadd_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-wasm-x1.c b/src/f32-vbinary/gen/vadd-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vadd-wasm-x1.c
rename to src/f32-vbinary/gen/vadd-minmax-wasm-x1.c
index 7bf784d..52b1122 100644
--- a/src/f32-vbinary/gen/vadd-wasm-x1.c
+++ b/src/f32-vbinary/gen/vadd-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__wasm_x1(
+void xnn_f32_vadd_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-wasm-x2.c b/src/f32-vbinary/gen/vadd-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vadd-wasm-x2.c
rename to src/f32-vbinary/gen/vadd-minmax-wasm-x2.c
index e4c7e7c..f1e7eb7 100644
--- a/src/f32-vbinary/gen/vadd-wasm-x2.c
+++ b/src/f32-vbinary/gen/vadd-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__wasm_x2(
+void xnn_f32_vadd_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vadd-wasm-x4.c b/src/f32-vbinary/gen/vadd-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vadd-wasm-x4.c
rename to src/f32-vbinary/gen/vadd-minmax-wasm-x4.c
index be5652a..a9ab718 100644
--- a/src/f32-vbinary/gen/vadd-wasm-x4.c
+++ b/src/f32-vbinary/gen/vadd-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vadd_ukernel__wasm_x4(
+void xnn_f32_vadd_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-avx-x16.c b/src/f32-vbinary/gen/vaddc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vaddc-avx-x16.c
rename to src/f32-vbinary/gen/vaddc-minmax-avx-x16.c
index 216a00c..db6ffb8 100644
--- a/src/f32-vbinary/gen/vaddc-avx-x16.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vaddc_ukernel__avx_x16(
+void xnn_f32_vaddc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-avx-x8.c b/src/f32-vbinary/gen/vaddc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-avx-x8.c
rename to src/f32-vbinary/gen/vaddc-minmax-avx-x8.c
index 2f6f4ca..cfb6eac 100644
--- a/src/f32-vbinary/gen/vaddc-avx-x8.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vaddc_ukernel__avx_x8(
+void xnn_f32_vaddc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-avx512f-x16.c b/src/f32-vbinary/gen/vaddc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-avx512f-x16.c
rename to src/f32-vbinary/gen/vaddc-minmax-avx512f-x16.c
index 3397651..a3ab579 100644
--- a/src/f32-vbinary/gen/vaddc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__avx512f_x16(
+void xnn_f32_vaddc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-avx512f-x32.c b/src/f32-vbinary/gen/vaddc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-avx512f-x32.c
rename to src/f32-vbinary/gen/vaddc-minmax-avx512f-x32.c
index a0d767c..c7cd3e3 100644
--- a/src/f32-vbinary/gen/vaddc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__avx512f_x32(
+void xnn_f32_vaddc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-neon-x4.c b/src/f32-vbinary/gen/vaddc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-neon-x4.c
rename to src/f32-vbinary/gen/vaddc-minmax-neon-x4.c
index 894b6d2..442420d 100644
--- a/src/f32-vbinary/gen/vaddc-neon-x4.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__neon_x4(
+void xnn_f32_vaddc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-neon-x8.c b/src/f32-vbinary/gen/vaddc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-neon-x8.c
rename to src/f32-vbinary/gen/vaddc-minmax-neon-x8.c
index 145506a..c99a103 100644
--- a/src/f32-vbinary/gen/vaddc-neon-x8.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__neon_x8(
+void xnn_f32_vaddc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-psimd-x4.c b/src/f32-vbinary/gen/vaddc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-psimd-x4.c
rename to src/f32-vbinary/gen/vaddc-minmax-psimd-x4.c
index a371b0e..2266ff3 100644
--- a/src/f32-vbinary/gen/vaddc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__psimd_x4(
+void xnn_f32_vaddc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-psimd-x8.c b/src/f32-vbinary/gen/vaddc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-psimd-x8.c
rename to src/f32-vbinary/gen/vaddc-minmax-psimd-x8.c
index c042a4c..2dd28df 100644
--- a/src/f32-vbinary/gen/vaddc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__psimd_x8(
+void xnn_f32_vaddc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-scalar-x1.c b/src/f32-vbinary/gen/vaddc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vaddc-scalar-x1.c
rename to src/f32-vbinary/gen/vaddc-minmax-scalar-x1.c
index 85a35b2..b424d51 100644
--- a/src/f32-vbinary/gen/vaddc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__scalar_x1(
+void xnn_f32_vaddc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-scalar-x2.c b/src/f32-vbinary/gen/vaddc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vaddc-scalar-x2.c
rename to src/f32-vbinary/gen/vaddc-minmax-scalar-x2.c
index dcb6de2..2958a84 100644
--- a/src/f32-vbinary/gen/vaddc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__scalar_x2(
+void xnn_f32_vaddc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-scalar-x4.c b/src/f32-vbinary/gen/vaddc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vaddc-scalar-x4.c
rename to src/f32-vbinary/gen/vaddc-minmax-scalar-x4.c
index 915ee50..61d297b 100644
--- a/src/f32-vbinary/gen/vaddc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__scalar_x4(
+void xnn_f32_vaddc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-sse-x4.c b/src/f32-vbinary/gen/vaddc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-sse-x4.c
rename to src/f32-vbinary/gen/vaddc-minmax-sse-x4.c
index 9408e81..e618ef3 100644
--- a/src/f32-vbinary/gen/vaddc-sse-x4.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__sse_x4(
+void xnn_f32_vaddc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-sse-x8.c b/src/f32-vbinary/gen/vaddc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-sse-x8.c
rename to src/f32-vbinary/gen/vaddc-minmax-sse-x8.c
index 5fb7072..c2b18b5 100644
--- a/src/f32-vbinary/gen/vaddc-sse-x8.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__sse_x8(
+void xnn_f32_vaddc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-wasm-x1.c b/src/f32-vbinary/gen/vaddc-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vaddc-wasm-x1.c
rename to src/f32-vbinary/gen/vaddc-minmax-wasm-x1.c
index 354ec47..5a86a29 100644
--- a/src/f32-vbinary/gen/vaddc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__wasm_x1(
+void xnn_f32_vaddc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-wasm-x2.c b/src/f32-vbinary/gen/vaddc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vaddc-wasm-x2.c
rename to src/f32-vbinary/gen/vaddc-minmax-wasm-x2.c
index 6f150fe..42af1e4 100644
--- a/src/f32-vbinary/gen/vaddc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__wasm_x2(
+void xnn_f32_vaddc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vaddc-wasm-x4.c b/src/f32-vbinary/gen/vaddc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vaddc-wasm-x4.c
rename to src/f32-vbinary/gen/vaddc-minmax-wasm-x4.c
index d881fae..142f7e0 100644
--- a/src/f32-vbinary/gen/vaddc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vaddc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vaddc_ukernel__wasm_x4(
+void xnn_f32_vaddc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-avx-x16.c b/src/f32-vbinary/gen/vdiv-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vdiv-avx-x16.c
rename to src/f32-vbinary/gen/vdiv-minmax-avx-x16.c
index b10a302..3150d56 100644
--- a/src/f32-vbinary/gen/vdiv-avx-x16.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vdiv_ukernel__avx_x16(
+void xnn_f32_vdiv_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-avx-x8.c b/src/f32-vbinary/gen/vdiv-minmax-avx-x8.c
similarity index 98%
rename from src/f32-vbinary/gen/vdiv-avx-x8.c
rename to src/f32-vbinary/gen/vdiv-minmax-avx-x8.c
index 8b868e1..44e231c 100644
--- a/src/f32-vbinary/gen/vdiv-avx-x8.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vdiv_ukernel__avx_x8(
+void xnn_f32_vdiv_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-avx512f-x16.c b/src/f32-vbinary/gen/vdiv-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-avx512f-x16.c
rename to src/f32-vbinary/gen/vdiv-minmax-avx512f-x16.c
index caf9c80..ba15e10 100644
--- a/src/f32-vbinary/gen/vdiv-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__avx512f_x16(
+void xnn_f32_vdiv_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-avx512f-x32.c b/src/f32-vbinary/gen/vdiv-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-avx512f-x32.c
rename to src/f32-vbinary/gen/vdiv-minmax-avx512f-x32.c
index 581e84c..afde483 100644
--- a/src/f32-vbinary/gen/vdiv-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__avx512f_x32(
+void xnn_f32_vdiv_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-neon-x4.c b/src/f32-vbinary/gen/vdiv-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-neon-x4.c
rename to src/f32-vbinary/gen/vdiv-minmax-neon-x4.c
index ab69795..bdabcca 100644
--- a/src/f32-vbinary/gen/vdiv-neon-x4.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__neon_x4(
+void xnn_f32_vdiv_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-neon-x8.c b/src/f32-vbinary/gen/vdiv-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-neon-x8.c
rename to src/f32-vbinary/gen/vdiv-minmax-neon-x8.c
index 7704334..6e00f72 100644
--- a/src/f32-vbinary/gen/vdiv-neon-x8.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__neon_x8(
+void xnn_f32_vdiv_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-psimd-x4.c b/src/f32-vbinary/gen/vdiv-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-psimd-x4.c
rename to src/f32-vbinary/gen/vdiv-minmax-psimd-x4.c
index f57a433..b27ecf0 100644
--- a/src/f32-vbinary/gen/vdiv-psimd-x4.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__psimd_x4(
+void xnn_f32_vdiv_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-psimd-x8.c b/src/f32-vbinary/gen/vdiv-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-psimd-x8.c
rename to src/f32-vbinary/gen/vdiv-minmax-psimd-x8.c
index 009285d..d82b3c1 100644
--- a/src/f32-vbinary/gen/vdiv-psimd-x8.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__psimd_x8(
+void xnn_f32_vdiv_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-scalar-x1.c b/src/f32-vbinary/gen/vdiv-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vdiv-scalar-x1.c
rename to src/f32-vbinary/gen/vdiv-minmax-scalar-x1.c
index fdc2485..a3dfd20 100644
--- a/src/f32-vbinary/gen/vdiv-scalar-x1.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__scalar_x1(
+void xnn_f32_vdiv_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-scalar-x2.c b/src/f32-vbinary/gen/vdiv-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vdiv-scalar-x2.c
rename to src/f32-vbinary/gen/vdiv-minmax-scalar-x2.c
index 65dd914..c9c5a4e 100644
--- a/src/f32-vbinary/gen/vdiv-scalar-x2.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__scalar_x2(
+void xnn_f32_vdiv_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-scalar-x4.c b/src/f32-vbinary/gen/vdiv-minmax-scalar-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-scalar-x4.c
rename to src/f32-vbinary/gen/vdiv-minmax-scalar-x4.c
index aa0f3a9..be08854 100644
--- a/src/f32-vbinary/gen/vdiv-scalar-x4.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__scalar_x4(
+void xnn_f32_vdiv_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-sse-x4.c b/src/f32-vbinary/gen/vdiv-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-sse-x4.c
rename to src/f32-vbinary/gen/vdiv-minmax-sse-x4.c
index f7a2ad4..ab481b0 100644
--- a/src/f32-vbinary/gen/vdiv-sse-x4.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__sse_x4(
+void xnn_f32_vdiv_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-sse-x8.c b/src/f32-vbinary/gen/vdiv-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-sse-x8.c
rename to src/f32-vbinary/gen/vdiv-minmax-sse-x8.c
index 99a8c34..564b48a 100644
--- a/src/f32-vbinary/gen/vdiv-sse-x8.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__sse_x8(
+void xnn_f32_vdiv_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-wasm-x1.c b/src/f32-vbinary/gen/vdiv-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vdiv-wasm-x1.c
rename to src/f32-vbinary/gen/vdiv-minmax-wasm-x1.c
index a64b993..9ad14c2 100644
--- a/src/f32-vbinary/gen/vdiv-wasm-x1.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__wasm_x1(
+void xnn_f32_vdiv_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-wasm-x2.c b/src/f32-vbinary/gen/vdiv-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vdiv-wasm-x2.c
rename to src/f32-vbinary/gen/vdiv-minmax-wasm-x2.c
index 2f607e1..cc530b3 100644
--- a/src/f32-vbinary/gen/vdiv-wasm-x2.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__wasm_x2(
+void xnn_f32_vdiv_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdiv-wasm-x4.c b/src/f32-vbinary/gen/vdiv-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdiv-wasm-x4.c
rename to src/f32-vbinary/gen/vdiv-minmax-wasm-x4.c
index 8ca9cc8..6f178b4 100644
--- a/src/f32-vbinary/gen/vdiv-wasm-x4.c
+++ b/src/f32-vbinary/gen/vdiv-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdiv_ukernel__wasm_x4(
+void xnn_f32_vdiv_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-avx-x16.c b/src/f32-vbinary/gen/vdivc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vdivc-avx-x16.c
rename to src/f32-vbinary/gen/vdivc-minmax-avx-x16.c
index a9ab3b6..fd72c61 100644
--- a/src/f32-vbinary/gen/vdivc-avx-x16.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vdivc_ukernel__avx_x16(
+void xnn_f32_vdivc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-avx-x8.c b/src/f32-vbinary/gen/vdivc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-avx-x8.c
rename to src/f32-vbinary/gen/vdivc-minmax-avx-x8.c
index 5411c8e..f5f88bb 100644
--- a/src/f32-vbinary/gen/vdivc-avx-x8.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vdivc_ukernel__avx_x8(
+void xnn_f32_vdivc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-avx512f-x16.c b/src/f32-vbinary/gen/vdivc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-avx512f-x16.c
rename to src/f32-vbinary/gen/vdivc-minmax-avx512f-x16.c
index 2608a12..64de6857 100644
--- a/src/f32-vbinary/gen/vdivc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__avx512f_x16(
+void xnn_f32_vdivc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-avx512f-x32.c b/src/f32-vbinary/gen/vdivc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-avx512f-x32.c
rename to src/f32-vbinary/gen/vdivc-minmax-avx512f-x32.c
index 96b457c..d00f88d 100644
--- a/src/f32-vbinary/gen/vdivc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__avx512f_x32(
+void xnn_f32_vdivc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-neon-x4.c b/src/f32-vbinary/gen/vdivc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-neon-x4.c
rename to src/f32-vbinary/gen/vdivc-minmax-neon-x4.c
index 133feed..488e4f6 100644
--- a/src/f32-vbinary/gen/vdivc-neon-x4.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__neon_x4(
+void xnn_f32_vdivc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-neon-x8.c b/src/f32-vbinary/gen/vdivc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-neon-x8.c
rename to src/f32-vbinary/gen/vdivc-minmax-neon-x8.c
index 226fe86..7b05df3 100644
--- a/src/f32-vbinary/gen/vdivc-neon-x8.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__neon_x8(
+void xnn_f32_vdivc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-psimd-x4.c b/src/f32-vbinary/gen/vdivc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-psimd-x4.c
rename to src/f32-vbinary/gen/vdivc-minmax-psimd-x4.c
index d6c1373..f09b98c 100644
--- a/src/f32-vbinary/gen/vdivc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__psimd_x4(
+void xnn_f32_vdivc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-psimd-x8.c b/src/f32-vbinary/gen/vdivc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-psimd-x8.c
rename to src/f32-vbinary/gen/vdivc-minmax-psimd-x8.c
index ca80f01..118b735 100644
--- a/src/f32-vbinary/gen/vdivc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__psimd_x8(
+void xnn_f32_vdivc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-scalar-x1.c b/src/f32-vbinary/gen/vdivc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vdivc-scalar-x1.c
rename to src/f32-vbinary/gen/vdivc-minmax-scalar-x1.c
index 576eba8..e780536 100644
--- a/src/f32-vbinary/gen/vdivc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__scalar_x1(
+void xnn_f32_vdivc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-scalar-x2.c b/src/f32-vbinary/gen/vdivc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vdivc-scalar-x2.c
rename to src/f32-vbinary/gen/vdivc-minmax-scalar-x2.c
index 7137e2f..6118d1b 100644
--- a/src/f32-vbinary/gen/vdivc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__scalar_x2(
+void xnn_f32_vdivc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-scalar-x4.c b/src/f32-vbinary/gen/vdivc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vdivc-scalar-x4.c
rename to src/f32-vbinary/gen/vdivc-minmax-scalar-x4.c
index e7cda2a..815fc6b 100644
--- a/src/f32-vbinary/gen/vdivc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__scalar_x4(
+void xnn_f32_vdivc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-sse-x4.c b/src/f32-vbinary/gen/vdivc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-sse-x4.c
rename to src/f32-vbinary/gen/vdivc-minmax-sse-x4.c
index cc27bcb..ad13aa7 100644
--- a/src/f32-vbinary/gen/vdivc-sse-x4.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__sse_x4(
+void xnn_f32_vdivc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-sse-x8.c b/src/f32-vbinary/gen/vdivc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-sse-x8.c
rename to src/f32-vbinary/gen/vdivc-minmax-sse-x8.c
index 96387ce..bce65b3 100644
--- a/src/f32-vbinary/gen/vdivc-sse-x8.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__sse_x8(
+void xnn_f32_vdivc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-wasm-x1.c b/src/f32-vbinary/gen/vdivc-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vdivc-wasm-x1.c
rename to src/f32-vbinary/gen/vdivc-minmax-wasm-x1.c
index fba49f9..b6d4efe 100644
--- a/src/f32-vbinary/gen/vdivc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__wasm_x1(
+void xnn_f32_vdivc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-wasm-x2.c b/src/f32-vbinary/gen/vdivc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vdivc-wasm-x2.c
rename to src/f32-vbinary/gen/vdivc-minmax-wasm-x2.c
index 614a0e9..0899037 100644
--- a/src/f32-vbinary/gen/vdivc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__wasm_x2(
+void xnn_f32_vdivc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vdivc-wasm-x4.c b/src/f32-vbinary/gen/vdivc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vdivc-wasm-x4.c
rename to src/f32-vbinary/gen/vdivc-minmax-wasm-x4.c
index 063ce25..0fa5050 100644
--- a/src/f32-vbinary/gen/vdivc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vdivc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vdivc_ukernel__wasm_x4(
+void xnn_f32_vdivc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmax-avx-x16.c b/src/f32-vbinary/gen/vmax-avx-x16.c
index 61a6e9c..ea6fc43 100644
--- a/src/f32-vbinary/gen/vmax-avx-x16.c
+++ b/src/f32-vbinary/gen/vmax-avx-x16.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
const __m256 va01234567 = _mm256_loadu_ps(a);
@@ -42,11 +40,6 @@
__m256 vy01234567 = _mm256_max_ps(va01234567, vb01234567);
__m256 vy89ABCDEF = _mm256_max_ps(va89ABCDEF, vb89ABCDEF);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
- vy89ABCDEF = _mm256_max_ps(vy89ABCDEF, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
- vy89ABCDEF = _mm256_min_ps(vy89ABCDEF, vy_max);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
@@ -60,8 +53,6 @@
b += 8;
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -74,8 +65,6 @@
const __m256 vb = _mm256_maskload_ps(b, vmask);
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmax-avx-x8.c b/src/f32-vbinary/gen/vmax-avx-x8.c
index 79adac3..97fa8b8 100644
--- a/src/f32-vbinary/gen/vmax-avx-x8.c
+++ b/src/f32-vbinary/gen/vmax-avx-x8.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const __m256 va01234567 = _mm256_loadu_ps(a);
@@ -39,9 +37,6 @@
__m256 vy01234567 = _mm256_max_ps(va01234567, vb01234567);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
_mm256_storeu_ps(y, vy01234567);
y += 8;
@@ -54,8 +49,6 @@
b += 8;
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -68,8 +61,6 @@
const __m256 vb = _mm256_maskload_ps(b, vmask);
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmax-avx512f-x16.c b/src/f32-vbinary/gen/vmax-avx512f-x16.c
index 96a565e..8ddee80 100644
--- a/src/f32-vbinary/gen/vmax-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vmax-avx512f-x16.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
const __m512 va0123456789ABCDEF = _mm512_loadu_ps(a);
@@ -38,9 +36,6 @@
__m512 vy0123456789ABCDEF = _mm512_max_ps(va0123456789ABCDEF, vb0123456789ABCDEF);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
y += 16;
@@ -53,8 +48,6 @@
b += 16;
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -69,8 +62,6 @@
const __m512 vb = _mm512_maskz_loadu_ps(vmask, b);
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmax-avx512f-x32.c b/src/f32-vbinary/gen/vmax-avx512f-x32.c
index 4dc8f88..aaa85d2 100644
--- a/src/f32-vbinary/gen/vmax-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vmax-avx512f-x32.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
for (; n >= 32 * sizeof(float); n -= 32 * sizeof(float)) {
const __m512 va0123456789ABCDEF = _mm512_loadu_ps(a);
@@ -41,11 +39,6 @@
__m512 vy0123456789ABCDEF = _mm512_max_ps(va0123456789ABCDEF, vb0123456789ABCDEF);
__m512 vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vaGHIJKLMNOPQRSTUV, vbGHIJKLMNOPQRSTUV);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
- vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vyGHIJKLMNOPQRSTUV, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
- vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vyGHIJKLMNOPQRSTUV, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
_mm512_storeu_ps(y + 16, vyGHIJKLMNOPQRSTUV);
@@ -59,8 +52,6 @@
b += 16;
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -75,8 +66,6 @@
const __m512 vb = _mm512_maskz_loadu_ps(vmask, b);
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmax-neon-x4.c b/src/f32-vbinary/gen/vmax-neon-x4.c
index 49a2fea..920d281 100644
--- a/src/f32-vbinary/gen/vmax-neon-x4.c
+++ b/src/f32-vbinary/gen/vmax-neon-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float32x4_t va0123 = vld1q_f32(a); a += 4;
@@ -34,9 +32,6 @@
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
@@ -45,8 +40,6 @@
const float32x4_t vb0123 = vld1q_f32(b); b += 4;
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
@@ -54,8 +47,6 @@
const float32x4_t vb0123 = vld1q_f32(b);
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmax-neon-x8.c b/src/f32-vbinary/gen/vmax-neon-x8.c
index 55aac28..3403389 100644
--- a/src/f32-vbinary/gen/vmax-neon-x8.c
+++ b/src/f32-vbinary/gen/vmax-neon-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const float32x4_t va0123 = vld1q_f32(a); a += 4;
@@ -37,11 +35,6 @@
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
float32x4_t vy4567 = vmaxq_f32(va4567, vb4567);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy4567 = vmaxq_f32(vy4567, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
- vy4567 = vminq_f32(vy4567, vy_max);
vst1q_f32(y, vy0123); y += 4;
vst1q_f32(y, vy4567); y += 4;
@@ -51,8 +44,6 @@
const float32x4_t vb0123 = vld1q_f32(b); b += 4;
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
@@ -60,8 +51,6 @@
const float32x4_t vb0123 = vld1q_f32(b);
float32x4_t vy0123 = vmaxq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmax-psimd-x4.c b/src/f32-vbinary/gen/vmax-psimd-x4.c
index 3fb08a0..704a279 100644
--- a/src/f32-vbinary/gen/vmax-psimd-x4.c
+++ b/src/f32-vbinary/gen/vmax-psimd-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const psimd_f32 va0123 = psimd_load_f32(a);
@@ -37,9 +35,6 @@
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
@@ -52,8 +47,6 @@
b += 4;
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -62,8 +55,6 @@
const psimd_f32 vb0123 = psimd_load_f32(b);
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmax-psimd-x8.c b/src/f32-vbinary/gen/vmax-psimd-x8.c
index 9d62229..35b439a 100644
--- a/src/f32-vbinary/gen/vmax-psimd-x8.c
+++ b/src/f32-vbinary/gen/vmax-psimd-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const psimd_f32 va0123 = psimd_load_f32(a);
@@ -40,11 +38,6 @@
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
psimd_f32 vy4567 = psimd_max_f32(va4567, vb4567);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy4567 = psimd_max_f32(vy4567, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
- vy4567 = psimd_min_f32(vy4567, vy_max);
psimd_store_f32(y, vy0123);
psimd_store_f32(y + 4, vy4567);
@@ -58,8 +51,6 @@
b += 4;
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -68,8 +59,6 @@
const psimd_f32 vb0123 = psimd_load_f32(b);
psimd_f32 vy0123 = psimd_max_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmax-scalar-x1.c b/src/f32-vbinary/gen/vmax-scalar-x1.c
index e85854c..54f7ed3 100644
--- a/src/f32-vbinary/gen/vmax-scalar-x1.c
+++ b/src/f32-vbinary/gen/vmax-scalar-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
const float vb = *b++;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmax-scalar-x2.c b/src/f32-vbinary/gen/vmax-scalar-x2.c
index 8d95310..7712735 100644
--- a/src/f32-vbinary/gen/vmax-scalar-x2.c
+++ b/src/f32-vbinary/gen/vmax-scalar-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
const float va0 = a[0];
@@ -39,11 +37,6 @@
float vy0 = math_max_f32(va0, vb0);
float vy1 = math_max_f32(va1, vb1);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -53,8 +46,6 @@
const float va = *a;
const float vb = *b;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmax-scalar-x4.c b/src/f32-vbinary/gen/vmax-scalar-x4.c
index df8a7a2..7dd203c 100644
--- a/src/f32-vbinary/gen/vmax-scalar-x4.c
+++ b/src/f32-vbinary/gen/vmax-scalar-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float va0 = a[0];
@@ -45,15 +43,6 @@
float vy2 = math_max_f32(va2, vb2);
float vy3 = math_max_f32(va3, vb3);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
- vy2 = math_max_f32(vy2, vy_min);
- vy3 = math_max_f32(vy3, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
- vy2 = math_min_f32(vy2, vy_max);
- vy3 = math_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -66,8 +55,6 @@
const float va = *a++;
const float vb = *b++;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmax-sse-x4.c b/src/f32-vbinary/gen/vmax-sse-x4.c
index aa37d68..b676f42 100644
--- a/src/f32-vbinary/gen/vmax-sse-x4.c
+++ b/src/f32-vbinary/gen/vmax-sse-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const __m128 va0123 = _mm_loadu_ps(a);
@@ -37,9 +35,6 @@
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
@@ -52,8 +47,6 @@
b += 4;
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -62,8 +55,6 @@
const __m128 vb0123 = _mm_loadu_ps(b);
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmax-sse-x8.c b/src/f32-vbinary/gen/vmax-sse-x8.c
index ef4bde3..b6a2e23 100644
--- a/src/f32-vbinary/gen/vmax-sse-x8.c
+++ b/src/f32-vbinary/gen/vmax-sse-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const __m128 va0123 = _mm_loadu_ps(a);
@@ -40,11 +38,6 @@
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
__m128 vy4567 = _mm_max_ps(va4567, vb4567);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy4567 = _mm_max_ps(vy4567, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
- vy4567 = _mm_min_ps(vy4567, vy_max);
_mm_storeu_ps(y, vy0123);
_mm_storeu_ps(y + 4, vy4567);
@@ -58,8 +51,6 @@
b += 4;
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -68,8 +59,6 @@
const __m128 vb0123 = _mm_loadu_ps(b);
__m128 vy0123 = _mm_max_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmax-wasm-x1.c b/src/f32-vbinary/gen/vmax-wasm-x1.c
index e0cc7ec..5ffa63e 100644
--- a/src/f32-vbinary/gen/vmax-wasm-x1.c
+++ b/src/f32-vbinary/gen/vmax-wasm-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
const float vb = *b++;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmax-wasm-x2.c b/src/f32-vbinary/gen/vmax-wasm-x2.c
index 85daf62..38f1df5 100644
--- a/src/f32-vbinary/gen/vmax-wasm-x2.c
+++ b/src/f32-vbinary/gen/vmax-wasm-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
const float va0 = a[0];
@@ -39,11 +37,6 @@
float vy0 = __builtin_wasm_max_f32(va0, vb0);
float vy1 = __builtin_wasm_max_f32(va1, vb1);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -53,8 +46,6 @@
const float va = *a;
const float vb = *b;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmax-wasm-x4.c b/src/f32-vbinary/gen/vmax-wasm-x4.c
index 4735b42..0b4c8ec 100644
--- a/src/f32-vbinary/gen/vmax-wasm-x4.c
+++ b/src/f32-vbinary/gen/vmax-wasm-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float va0 = a[0];
@@ -45,15 +43,6 @@
float vy2 = __builtin_wasm_max_f32(va2, vb2);
float vy3 = __builtin_wasm_max_f32(va3, vb3);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
- vy2 = __builtin_wasm_max_f32(vy2, vy_min);
- vy3 = __builtin_wasm_max_f32(vy3, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
- vy2 = __builtin_wasm_min_f32(vy2, vy_max);
- vy3 = __builtin_wasm_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -66,8 +55,6 @@
const float va = *a++;
const float vb = *b++;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmaxc-avx-x16.c b/src/f32-vbinary/gen/vmaxc-avx-x16.c
index 3d7d7fb..f6797e3 100644
--- a/src/f32-vbinary/gen/vmaxc-avx-x16.c
+++ b/src/f32-vbinary/gen/vmaxc-avx-x16.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
const __m256 vb = _mm256_broadcast_ss(b);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
@@ -39,11 +37,6 @@
__m256 vy01234567 = _mm256_max_ps(va01234567, vb);
__m256 vy89ABCDEF = _mm256_max_ps(va89ABCDEF, vb);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
- vy89ABCDEF = _mm256_max_ps(vy89ABCDEF, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
- vy89ABCDEF = _mm256_min_ps(vy89ABCDEF, vy_max);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
@@ -54,8 +47,6 @@
a += 8;
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -67,8 +58,6 @@
const __m256 va = _mm256_maskload_ps(a, vmask);
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmaxc-avx-x8.c b/src/f32-vbinary/gen/vmaxc-avx-x8.c
index cc0ad5d..9488d5c 100644
--- a/src/f32-vbinary/gen/vmaxc-avx-x8.c
+++ b/src/f32-vbinary/gen/vmaxc-avx-x8.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
const __m256 vb = _mm256_broadcast_ss(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,9 +35,6 @@
__m256 vy01234567 = _mm256_max_ps(va01234567, vb);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
_mm256_storeu_ps(y, vy01234567);
y += 8;
@@ -49,8 +44,6 @@
a += 8;
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -62,8 +55,6 @@
const __m256 va = _mm256_maskload_ps(a, vmask);
__m256 vy = _mm256_max_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmaxc-avx512f-x16.c b/src/f32-vbinary/gen/vmaxc-avx512f-x16.c
index 065eb85..fd6624c 100644
--- a/src/f32-vbinary/gen/vmaxc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vmaxc-avx512f-x16.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
const __m512 vb = _mm512_set1_ps(*b);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
@@ -36,9 +34,6 @@
__m512 vy0123456789ABCDEF = _mm512_max_ps(va0123456789ABCDEF, vb);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
y += 16;
@@ -48,8 +43,6 @@
a += 16;
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -63,8 +56,6 @@
const __m512 va = _mm512_maskz_loadu_ps(vmask, a);
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-avx512f-x32.c b/src/f32-vbinary/gen/vmaxc-avx512f-x32.c
index b753714..c8f5acf 100644
--- a/src/f32-vbinary/gen/vmaxc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vmaxc-avx512f-x32.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
const __m512 vb = _mm512_set1_ps(*b);
for (; n >= 32 * sizeof(float); n -= 32 * sizeof(float)) {
@@ -38,11 +36,6 @@
__m512 vy0123456789ABCDEF = _mm512_max_ps(va0123456789ABCDEF, vb);
__m512 vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vaGHIJKLMNOPQRSTUV, vb);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
- vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vyGHIJKLMNOPQRSTUV, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
- vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vyGHIJKLMNOPQRSTUV, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
_mm512_storeu_ps(y + 16, vyGHIJKLMNOPQRSTUV);
@@ -53,8 +46,6 @@
a += 16;
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -68,8 +59,6 @@
const __m512 va = _mm512_maskz_loadu_ps(vmask, a);
__m512 vy = _mm512_max_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-neon-x4.c b/src/f32-vbinary/gen/vmaxc-neon-x4.c
index 7907fdd..f0d31a9 100644
--- a/src/f32-vbinary/gen/vmaxc-neon-x4.c
+++ b/src/f32-vbinary/gen/vmaxc-neon-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
const float32x4_t vb = vld1q_dup_f32(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -34,9 +32,6 @@
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
@@ -44,16 +39,12 @@
const float32x4_t va0123 = vld1q_f32(a); a += 4;
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
const float32x4_t va0123 = vld1q_f32(a);
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmaxc-neon-x8.c b/src/f32-vbinary/gen/vmaxc-neon-x8.c
index c06875c..3ae9776 100644
--- a/src/f32-vbinary/gen/vmaxc-neon-x8.c
+++ b/src/f32-vbinary/gen/vmaxc-neon-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
const float32x4_t vb = vld1q_dup_f32(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -36,11 +34,6 @@
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
float32x4_t vy4567 = vmaxq_f32(va4567, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy4567 = vmaxq_f32(vy4567, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
- vy4567 = vminq_f32(vy4567, vy_max);
vst1q_f32(y, vy0123); y += 4;
vst1q_f32(y, vy4567); y += 4;
@@ -49,16 +42,12 @@
const float32x4_t va0123 = vld1q_f32(a); a += 4;
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
const float32x4_t va0123 = vld1q_f32(a);
float32x4_t vy0123 = vmaxq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmaxc-psimd-x4.c b/src/f32-vbinary/gen/vmaxc-psimd-x4.c
index 34abedf..d720c19 100644
--- a/src/f32-vbinary/gen/vmaxc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vmaxc-psimd-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
const psimd_f32 vb = psimd_load_splat_f32(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -35,9 +33,6 @@
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
@@ -47,8 +42,6 @@
a += 4;
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -56,8 +49,6 @@
const psimd_f32 va0123 = psimd_load_f32(a);
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmaxc-psimd-x8.c b/src/f32-vbinary/gen/vmaxc-psimd-x8.c
index 18a8f1c..aea9e1c 100644
--- a/src/f32-vbinary/gen/vmaxc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vmaxc-psimd-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
const psimd_f32 vb = psimd_load_splat_f32(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,11 +35,6 @@
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
psimd_f32 vy4567 = psimd_max_f32(va4567, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy4567 = psimd_max_f32(vy4567, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
- vy4567 = psimd_min_f32(vy4567, vy_max);
psimd_store_f32(y, vy0123);
psimd_store_f32(y + 4, vy4567);
@@ -52,8 +45,6 @@
a += 4;
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -61,8 +52,6 @@
const psimd_f32 va0123 = psimd_load_f32(a);
psimd_f32 vy0123 = psimd_max_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmaxc-scalar-x1.c b/src/f32-vbinary/gen/vmaxc-scalar-x1.c
index cd01cce..0ddc681 100644
--- a/src/f32-vbinary/gen/vmaxc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vmaxc-scalar-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-scalar-x2.c b/src/f32-vbinary/gen/vmaxc-scalar-x2.c
index 591f85a..7cb8550 100644
--- a/src/f32-vbinary/gen/vmaxc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vmaxc-scalar-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
@@ -36,11 +34,6 @@
float vy0 = math_max_f32(va0, vb);
float vy1 = math_max_f32(va1, vb);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -49,8 +42,6 @@
if XNN_UNLIKELY(n != 0) {
const float va = *a;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-scalar-x4.c b/src/f32-vbinary/gen/vmaxc-scalar-x4.c
index db3f2dd..fb04b69 100644
--- a/src/f32-vbinary/gen/vmaxc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vmaxc-scalar-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -40,15 +38,6 @@
float vy2 = math_max_f32(va2, vb);
float vy3 = math_max_f32(va3, vb);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
- vy2 = math_max_f32(vy2, vy_min);
- vy3 = math_max_f32(vy3, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
- vy2 = math_min_f32(vy2, vy_max);
- vy3 = math_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -60,8 +49,6 @@
do {
const float va = *a++;
float vy = math_max_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmaxc-sse-x4.c b/src/f32-vbinary/gen/vmaxc-sse-x4.c
index c082b71..b9a3c85 100644
--- a/src/f32-vbinary/gen/vmaxc-sse-x4.c
+++ b/src/f32-vbinary/gen/vmaxc-sse-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
const __m128 vb = _mm_load1_ps(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -35,9 +33,6 @@
__m128 vy0123 = _mm_max_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
@@ -47,8 +42,6 @@
a += 4;
__m128 vy0123 = _mm_max_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -56,8 +49,6 @@
const __m128 va0123 = _mm_loadu_ps(a);
__m128 vy0123 = _mm_max_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmaxc-sse-x8.c b/src/f32-vbinary/gen/vmaxc-sse-x8.c
index 10ed955..ef38289 100644
--- a/src/f32-vbinary/gen/vmaxc-sse-x8.c
+++ b/src/f32-vbinary/gen/vmaxc-sse-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
const __m128 vb = _mm_load1_ps(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,11 +35,6 @@
__m128 vy0123 = _mm_max_ps(va0123, vb);
__m128 vy4567 = _mm_max_ps(va4567, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy4567 = _mm_max_ps(vy4567, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
- vy4567 = _mm_min_ps(vy4567, vy_max);
_mm_storeu_ps(y, vy0123);
_mm_storeu_ps(y + 4, vy4567);
@@ -52,8 +45,6 @@
a += 4;
__m128 vy0123 = _mm_max_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -61,8 +52,6 @@
const __m128 va0123 = _mm_loadu_ps(a);
__m128 vy0123 = _mm_max_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmaxc-wasm-x1.c b/src/f32-vbinary/gen/vmaxc-wasm-x1.c
index 92d2181..3ac5488 100644
--- a/src/f32-vbinary/gen/vmaxc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vmaxc-wasm-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-wasm-x2.c b/src/f32-vbinary/gen/vmaxc-wasm-x2.c
index e0d8a5e..3741206 100644
--- a/src/f32-vbinary/gen/vmaxc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vmaxc-wasm-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
@@ -36,11 +34,6 @@
float vy0 = __builtin_wasm_max_f32(va0, vb);
float vy1 = __builtin_wasm_max_f32(va1, vb);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -49,8 +42,6 @@
if XNN_UNLIKELY(n != 0) {
const float va = *a;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmaxc-wasm-x4.c b/src/f32-vbinary/gen/vmaxc-wasm-x4.c
index fc4a6e9..5a78b43 100644
--- a/src/f32-vbinary/gen/vmaxc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vmaxc-wasm-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -40,15 +38,6 @@
float vy2 = __builtin_wasm_max_f32(va2, vb);
float vy3 = __builtin_wasm_max_f32(va3, vb);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
- vy2 = __builtin_wasm_max_f32(vy2, vy_min);
- vy3 = __builtin_wasm_max_f32(vy3, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
- vy2 = __builtin_wasm_min_f32(vy2, vy_max);
- vy3 = __builtin_wasm_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -60,8 +49,6 @@
do {
const float va = *a++;
float vy = __builtin_wasm_max_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmin-avx-x16.c b/src/f32-vbinary/gen/vmin-avx-x16.c
index 38603c2..dc4938c 100644
--- a/src/f32-vbinary/gen/vmin-avx-x16.c
+++ b/src/f32-vbinary/gen/vmin-avx-x16.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
const __m256 va01234567 = _mm256_loadu_ps(a);
@@ -42,11 +40,6 @@
__m256 vy01234567 = _mm256_min_ps(va01234567, vb01234567);
__m256 vy89ABCDEF = _mm256_min_ps(va89ABCDEF, vb89ABCDEF);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
- vy89ABCDEF = _mm256_max_ps(vy89ABCDEF, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
- vy89ABCDEF = _mm256_min_ps(vy89ABCDEF, vy_max);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
@@ -60,8 +53,6 @@
b += 8;
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -74,8 +65,6 @@
const __m256 vb = _mm256_maskload_ps(b, vmask);
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmin-avx-x8.c b/src/f32-vbinary/gen/vmin-avx-x8.c
index e0f52d0..7d057f2 100644
--- a/src/f32-vbinary/gen/vmin-avx-x8.c
+++ b/src/f32-vbinary/gen/vmin-avx-x8.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const __m256 va01234567 = _mm256_loadu_ps(a);
@@ -39,9 +37,6 @@
__m256 vy01234567 = _mm256_min_ps(va01234567, vb01234567);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
_mm256_storeu_ps(y, vy01234567);
y += 8;
@@ -54,8 +49,6 @@
b += 8;
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -68,8 +61,6 @@
const __m256 vb = _mm256_maskload_ps(b, vmask);
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vmin-avx512f-x16.c b/src/f32-vbinary/gen/vmin-avx512f-x16.c
index d01eed8..e7e3660 100644
--- a/src/f32-vbinary/gen/vmin-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vmin-avx512f-x16.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
const __m512 va0123456789ABCDEF = _mm512_loadu_ps(a);
@@ -38,9 +36,6 @@
__m512 vy0123456789ABCDEF = _mm512_min_ps(va0123456789ABCDEF, vb0123456789ABCDEF);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
y += 16;
@@ -53,8 +48,6 @@
b += 16;
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -69,8 +62,6 @@
const __m512 vb = _mm512_maskz_loadu_ps(vmask, b);
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmin-avx512f-x32.c b/src/f32-vbinary/gen/vmin-avx512f-x32.c
index 6ca07c9..f7520f6 100644
--- a/src/f32-vbinary/gen/vmin-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vmin-avx512f-x32.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
for (; n >= 32 * sizeof(float); n -= 32 * sizeof(float)) {
const __m512 va0123456789ABCDEF = _mm512_loadu_ps(a);
@@ -41,11 +39,6 @@
__m512 vy0123456789ABCDEF = _mm512_min_ps(va0123456789ABCDEF, vb0123456789ABCDEF);
__m512 vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vaGHIJKLMNOPQRSTUV, vbGHIJKLMNOPQRSTUV);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
- vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vyGHIJKLMNOPQRSTUV, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
- vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vyGHIJKLMNOPQRSTUV, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
_mm512_storeu_ps(y + 16, vyGHIJKLMNOPQRSTUV);
@@ -59,8 +52,6 @@
b += 16;
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -75,8 +66,6 @@
const __m512 vb = _mm512_maskz_loadu_ps(vmask, b);
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vmin-neon-x4.c b/src/f32-vbinary/gen/vmin-neon-x4.c
index 8f170a0..7243808 100644
--- a/src/f32-vbinary/gen/vmin-neon-x4.c
+++ b/src/f32-vbinary/gen/vmin-neon-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float32x4_t va0123 = vld1q_f32(a); a += 4;
@@ -34,9 +32,6 @@
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
@@ -45,8 +40,6 @@
const float32x4_t vb0123 = vld1q_f32(b); b += 4;
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
@@ -54,8 +47,6 @@
const float32x4_t vb0123 = vld1q_f32(b);
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmin-neon-x8.c b/src/f32-vbinary/gen/vmin-neon-x8.c
index 3dde9fe..d07d64f 100644
--- a/src/f32-vbinary/gen/vmin-neon-x8.c
+++ b/src/f32-vbinary/gen/vmin-neon-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const float32x4_t va0123 = vld1q_f32(a); a += 4;
@@ -37,11 +35,6 @@
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
float32x4_t vy4567 = vminq_f32(va4567, vb4567);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy4567 = vmaxq_f32(vy4567, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
- vy4567 = vminq_f32(vy4567, vy_max);
vst1q_f32(y, vy0123); y += 4;
vst1q_f32(y, vy4567); y += 4;
@@ -51,8 +44,6 @@
const float32x4_t vb0123 = vld1q_f32(b); b += 4;
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
@@ -60,8 +51,6 @@
const float32x4_t vb0123 = vld1q_f32(b);
float32x4_t vy0123 = vminq_f32(va0123, vb0123);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vmin-psimd-x4.c b/src/f32-vbinary/gen/vmin-psimd-x4.c
index d5a52d1..5821bb8 100644
--- a/src/f32-vbinary/gen/vmin-psimd-x4.c
+++ b/src/f32-vbinary/gen/vmin-psimd-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const psimd_f32 va0123 = psimd_load_f32(a);
@@ -37,9 +35,6 @@
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
@@ -52,8 +47,6 @@
b += 4;
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -62,8 +55,6 @@
const psimd_f32 vb0123 = psimd_load_f32(b);
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmin-psimd-x8.c b/src/f32-vbinary/gen/vmin-psimd-x8.c
index 451424e..4887b8e 100644
--- a/src/f32-vbinary/gen/vmin-psimd-x8.c
+++ b/src/f32-vbinary/gen/vmin-psimd-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const psimd_f32 va0123 = psimd_load_f32(a);
@@ -40,11 +38,6 @@
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
psimd_f32 vy4567 = psimd_min_f32(va4567, vb4567);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy4567 = psimd_max_f32(vy4567, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
- vy4567 = psimd_min_f32(vy4567, vy_max);
psimd_store_f32(y, vy0123);
psimd_store_f32(y + 4, vy4567);
@@ -58,8 +51,6 @@
b += 4;
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -68,8 +59,6 @@
const psimd_f32 vb0123 = psimd_load_f32(b);
psimd_f32 vy0123 = psimd_min_f32(va0123, vb0123);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmin-scalar-x1.c b/src/f32-vbinary/gen/vmin-scalar-x1.c
index 8c39b89..1716792 100644
--- a/src/f32-vbinary/gen/vmin-scalar-x1.c
+++ b/src/f32-vbinary/gen/vmin-scalar-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
const float vb = *b++;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmin-scalar-x2.c b/src/f32-vbinary/gen/vmin-scalar-x2.c
index fb789a2..1246c29 100644
--- a/src/f32-vbinary/gen/vmin-scalar-x2.c
+++ b/src/f32-vbinary/gen/vmin-scalar-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
const float va0 = a[0];
@@ -39,11 +37,6 @@
float vy0 = math_min_f32(va0, vb0);
float vy1 = math_min_f32(va1, vb1);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -53,8 +46,6 @@
const float va = *a;
const float vb = *b;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmin-scalar-x4.c b/src/f32-vbinary/gen/vmin-scalar-x4.c
index ba4ee2c..270809e 100644
--- a/src/f32-vbinary/gen/vmin-scalar-x4.c
+++ b/src/f32-vbinary/gen/vmin-scalar-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float va0 = a[0];
@@ -45,15 +43,6 @@
float vy2 = math_min_f32(va2, vb2);
float vy3 = math_min_f32(va3, vb3);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
- vy2 = math_max_f32(vy2, vy_min);
- vy3 = math_max_f32(vy3, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
- vy2 = math_min_f32(vy2, vy_max);
- vy3 = math_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -66,8 +55,6 @@
const float va = *a++;
const float vb = *b++;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmin-sse-x4.c b/src/f32-vbinary/gen/vmin-sse-x4.c
index 758efdb..948dafb 100644
--- a/src/f32-vbinary/gen/vmin-sse-x4.c
+++ b/src/f32-vbinary/gen/vmin-sse-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const __m128 va0123 = _mm_loadu_ps(a);
@@ -37,9 +35,6 @@
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
@@ -52,8 +47,6 @@
b += 4;
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -62,8 +55,6 @@
const __m128 vb0123 = _mm_loadu_ps(b);
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmin-sse-x8.c b/src/f32-vbinary/gen/vmin-sse-x8.c
index 285a039..0d5c866 100644
--- a/src/f32-vbinary/gen/vmin-sse-x8.c
+++ b/src/f32-vbinary/gen/vmin-sse-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const __m128 va0123 = _mm_loadu_ps(a);
@@ -40,11 +38,6 @@
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
__m128 vy4567 = _mm_min_ps(va4567, vb4567);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy4567 = _mm_max_ps(vy4567, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
- vy4567 = _mm_min_ps(vy4567, vy_max);
_mm_storeu_ps(y, vy0123);
_mm_storeu_ps(y + 4, vy4567);
@@ -58,8 +51,6 @@
b += 4;
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -68,8 +59,6 @@
const __m128 vb0123 = _mm_loadu_ps(b);
__m128 vy0123 = _mm_min_ps(va0123, vb0123);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vmin-wasm-x1.c b/src/f32-vbinary/gen/vmin-wasm-x1.c
index 3aa4b06..0e5dc8f 100644
--- a/src/f32-vbinary/gen/vmin-wasm-x1.c
+++ b/src/f32-vbinary/gen/vmin-wasm-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
const float vb = *b++;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmin-wasm-x2.c b/src/f32-vbinary/gen/vmin-wasm-x2.c
index d86a21a..fa21510 100644
--- a/src/f32-vbinary/gen/vmin-wasm-x2.c
+++ b/src/f32-vbinary/gen/vmin-wasm-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
const float va0 = a[0];
@@ -39,11 +37,6 @@
float vy0 = __builtin_wasm_min_f32(va0, vb0);
float vy1 = __builtin_wasm_min_f32(va1, vb1);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -53,8 +46,6 @@
const float va = *a;
const float vb = *b;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vmin-wasm-x4.c b/src/f32-vbinary/gen/vmin-wasm-x4.c
index 004bd12..b44a39e 100644
--- a/src/f32-vbinary/gen/vmin-wasm-x4.c
+++ b/src/f32-vbinary/gen/vmin-wasm-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
const float va0 = a[0];
@@ -45,15 +43,6 @@
float vy2 = __builtin_wasm_min_f32(va2, vb2);
float vy3 = __builtin_wasm_min_f32(va3, vb3);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
- vy2 = __builtin_wasm_max_f32(vy2, vy_min);
- vy3 = __builtin_wasm_max_f32(vy3, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
- vy2 = __builtin_wasm_min_f32(vy2, vy_max);
- vy3 = __builtin_wasm_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -66,8 +55,6 @@
const float va = *a++;
const float vb = *b++;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vminc-avx-x16.c b/src/f32-vbinary/gen/vminc-avx-x16.c
index 6b42b69..f69c66a 100644
--- a/src/f32-vbinary/gen/vminc-avx-x16.c
+++ b/src/f32-vbinary/gen/vminc-avx-x16.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
const __m256 vb = _mm256_broadcast_ss(b);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
@@ -39,11 +37,6 @@
__m256 vy01234567 = _mm256_min_ps(va01234567, vb);
__m256 vy89ABCDEF = _mm256_min_ps(va89ABCDEF, vb);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
- vy89ABCDEF = _mm256_max_ps(vy89ABCDEF, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
- vy89ABCDEF = _mm256_min_ps(vy89ABCDEF, vy_max);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
@@ -54,8 +47,6 @@
a += 8;
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -67,8 +58,6 @@
const __m256 va = _mm256_maskload_ps(a, vmask);
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vminc-avx-x8.c b/src/f32-vbinary/gen/vminc-avx-x8.c
index edfee93..9343a6e 100644
--- a/src/f32-vbinary/gen/vminc-avx-x8.c
+++ b/src/f32-vbinary/gen/vminc-avx-x8.c
@@ -22,13 +22,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m256 vy_min = _mm256_broadcast_ps((const __m128*) params->sse.min);
- const __m256 vy_max = _mm256_broadcast_ps((const __m128*) params->sse.max);
const __m256 vb = _mm256_broadcast_ss(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,9 +35,6 @@
__m256 vy01234567 = _mm256_min_ps(va01234567, vb);
- vy01234567 = _mm256_max_ps(vy01234567, vy_min);
-
- vy01234567 = _mm256_min_ps(vy01234567, vy_max);
_mm256_storeu_ps(y, vy01234567);
y += 8;
@@ -49,8 +44,6 @@
a += 8;
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
_mm256_storeu_ps(y, vy);
y += 8;
}
@@ -62,8 +55,6 @@
const __m256 va = _mm256_maskload_ps(a, vmask);
__m256 vy = _mm256_min_ps(va, vb);
- vy = _mm256_max_ps(vy, vy_min);
- vy = _mm256_min_ps(vy, vy_max);
// _mm256_maskstore_ps(y, vmask, vy) could be used here, but triggers msan failures (probably an msan bug).
__m128 vy_lo = _mm256_castps256_ps128(vy);
diff --git a/src/f32-vbinary/gen/vminc-avx512f-x16.c b/src/f32-vbinary/gen/vminc-avx512f-x16.c
index 9424f7b..13b0446 100644
--- a/src/f32-vbinary/gen/vminc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vminc-avx512f-x16.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
const __m512 vb = _mm512_set1_ps(*b);
for (; n >= 16 * sizeof(float); n -= 16 * sizeof(float)) {
@@ -36,9 +34,6 @@
__m512 vy0123456789ABCDEF = _mm512_min_ps(va0123456789ABCDEF, vb);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
y += 16;
@@ -48,8 +43,6 @@
a += 16;
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -63,8 +56,6 @@
const __m512 va = _mm512_maskz_loadu_ps(vmask, a);
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vminc-avx512f-x32.c b/src/f32-vbinary/gen/vminc-avx512f-x32.c
index de99bc8..a16e624 100644
--- a/src/f32-vbinary/gen/vminc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vminc-avx512f-x32.c
@@ -21,13 +21,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m512 vy_min = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.min));
- const __m512 vy_max = _mm512_broadcast_f32x4(_mm_load_ps(params->sse.max));
const __m512 vb = _mm512_set1_ps(*b);
for (; n >= 32 * sizeof(float); n -= 32 * sizeof(float)) {
@@ -38,11 +36,6 @@
__m512 vy0123456789ABCDEF = _mm512_min_ps(va0123456789ABCDEF, vb);
__m512 vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vaGHIJKLMNOPQRSTUV, vb);
- vy0123456789ABCDEF = _mm512_max_ps(vy0123456789ABCDEF, vy_min);
- vyGHIJKLMNOPQRSTUV = _mm512_max_ps(vyGHIJKLMNOPQRSTUV, vy_min);
-
- vy0123456789ABCDEF = _mm512_min_ps(vy0123456789ABCDEF, vy_max);
- vyGHIJKLMNOPQRSTUV = _mm512_min_ps(vyGHIJKLMNOPQRSTUV, vy_max);
_mm512_storeu_ps(y, vy0123456789ABCDEF);
_mm512_storeu_ps(y + 16, vyGHIJKLMNOPQRSTUV);
@@ -53,8 +46,6 @@
a += 16;
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_storeu_ps(y, vy);
y += 16;
}
@@ -68,8 +59,6 @@
const __m512 va = _mm512_maskz_loadu_ps(vmask, a);
__m512 vy = _mm512_min_ps(va, vb);
- vy = _mm512_max_ps(vy, vy_min);
- vy = _mm512_min_ps(vy, vy_max);
_mm512_mask_storeu_ps(y, vmask, vy);
}
}
diff --git a/src/f32-vbinary/gen/vminc-neon-x4.c b/src/f32-vbinary/gen/vminc-neon-x4.c
index 855f067..2927cd3 100644
--- a/src/f32-vbinary/gen/vminc-neon-x4.c
+++ b/src/f32-vbinary/gen/vminc-neon-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
const float32x4_t vb = vld1q_dup_f32(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -34,9 +32,6 @@
float32x4_t vy0123 = vminq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
@@ -44,16 +39,12 @@
const float32x4_t va0123 = vld1q_f32(a); a += 4;
float32x4_t vy0123 = vminq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
const float32x4_t va0123 = vld1q_f32(a);
float32x4_t vy0123 = vminq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vminc-neon-x8.c b/src/f32-vbinary/gen/vminc-neon-x8.c
index 92ac01f..c7d153a 100644
--- a/src/f32-vbinary/gen/vminc-neon-x8.c
+++ b/src/f32-vbinary/gen/vminc-neon-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float32x4_t vy_min = vld1q_dup_f32(¶ms->scalar.min);
- const float32x4_t vy_max = vld1q_dup_f32(¶ms->scalar.max);
const float32x4_t vb = vld1q_dup_f32(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -36,11 +34,6 @@
float32x4_t vy0123 = vminq_f32(va0123, vb);
float32x4_t vy4567 = vminq_f32(va4567, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy4567 = vmaxq_f32(vy4567, vy_min);
-
- vy0123 = vminq_f32(vy0123, vy_max);
- vy4567 = vminq_f32(vy4567, vy_max);
vst1q_f32(y, vy0123); y += 4;
vst1q_f32(y, vy4567); y += 4;
@@ -49,16 +42,12 @@
const float32x4_t va0123 = vld1q_f32(a); a += 4;
float32x4_t vy0123 = vminq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
vst1q_f32(y, vy0123); y += 4;
}
if XNN_UNLIKELY(n != 0) {
const float32x4_t va0123 = vld1q_f32(a);
float32x4_t vy0123 = vminq_f32(va0123, vb);
- vy0123 = vmaxq_f32(vy0123, vy_min);
- vy0123 = vminq_f32(vy0123, vy_max);
float32x2_t vy01 = vget_low_f32(vy0123);
if (n & (2 * sizeof(float))) {
diff --git a/src/f32-vbinary/gen/vminc-psimd-x4.c b/src/f32-vbinary/gen/vminc-psimd-x4.c
index 0553bd8..a11a7e5 100644
--- a/src/f32-vbinary/gen/vminc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vminc-psimd-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
const psimd_f32 vb = psimd_load_splat_f32(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -35,9 +33,6 @@
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
@@ -47,8 +42,6 @@
a += 4;
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -56,8 +49,6 @@
const psimd_f32 va0123 = psimd_load_f32(a);
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vminc-psimd-x8.c b/src/f32-vbinary/gen/vminc-psimd-x8.c
index 8746a7f..4866308 100644
--- a/src/f32-vbinary/gen/vminc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vminc-psimd-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const psimd_f32 vy_min = psimd_load_splat_f32(¶ms->scalar.min);
- const psimd_f32 vy_max = psimd_load_splat_f32(¶ms->scalar.max);
const psimd_f32 vb = psimd_load_splat_f32(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,11 +35,6 @@
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
psimd_f32 vy4567 = psimd_min_f32(va4567, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy4567 = psimd_max_f32(vy4567, vy_min);
-
- vy0123 = psimd_min_f32(vy0123, vy_max);
- vy4567 = psimd_min_f32(vy4567, vy_max);
psimd_store_f32(y, vy0123);
psimd_store_f32(y + 4, vy4567);
@@ -52,8 +45,6 @@
a += 4;
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
psimd_store_f32(y, vy0123);
y += 4;
}
@@ -61,8 +52,6 @@
const psimd_f32 va0123 = psimd_load_f32(a);
psimd_f32 vy0123 = psimd_min_f32(va0123, vb);
- vy0123 = psimd_max_f32(vy0123, vy_min);
- vy0123 = psimd_min_f32(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
psimd_store2_f32(y, vy0123);
vy0123 = psimd_concat_hi_f32(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vminc-scalar-x1.c b/src/f32-vbinary/gen/vminc-scalar-x1.c
index ccd33b6..b8aeab8 100644
--- a/src/f32-vbinary/gen/vminc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vminc-scalar-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vminc-scalar-x2.c b/src/f32-vbinary/gen/vminc-scalar-x2.c
index a0a6e08..07c4cc4 100644
--- a/src/f32-vbinary/gen/vminc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vminc-scalar-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
@@ -36,11 +34,6 @@
float vy0 = math_min_f32(va0, vb);
float vy1 = math_min_f32(va1, vb);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -49,8 +42,6 @@
if XNN_UNLIKELY(n != 0) {
const float va = *a;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vminc-scalar-x4.c b/src/f32-vbinary/gen/vminc-scalar-x4.c
index b16a77b..9f5a38a 100644
--- a/src/f32-vbinary/gen/vminc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vminc-scalar-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -40,15 +38,6 @@
float vy2 = math_min_f32(va2, vb);
float vy3 = math_min_f32(va3, vb);
- vy0 = math_max_f32(vy0, vy_min);
- vy1 = math_max_f32(vy1, vy_min);
- vy2 = math_max_f32(vy2, vy_min);
- vy3 = math_max_f32(vy3, vy_min);
-
- vy0 = math_min_f32(vy0, vy_max);
- vy1 = math_min_f32(vy1, vy_max);
- vy2 = math_min_f32(vy2, vy_max);
- vy3 = math_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -60,8 +49,6 @@
do {
const float va = *a++;
float vy = math_min_f32(va, vb);
- vy = math_max_f32(vy, vy_min);
- vy = math_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vminc-sse-x4.c b/src/f32-vbinary/gen/vminc-sse-x4.c
index 5f0d97c..f0ff610 100644
--- a/src/f32-vbinary/gen/vminc-sse-x4.c
+++ b/src/f32-vbinary/gen/vminc-sse-x4.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
const __m128 vb = _mm_load1_ps(b);
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -35,9 +33,6 @@
__m128 vy0123 = _mm_min_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
@@ -47,8 +42,6 @@
a += 4;
__m128 vy0123 = _mm_min_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -56,8 +49,6 @@
const __m128 va0123 = _mm_loadu_ps(a);
__m128 vy0123 = _mm_min_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vminc-sse-x8.c b/src/f32-vbinary/gen/vminc-sse-x8.c
index 9736006..7af7e20 100644
--- a/src/f32-vbinary/gen/vminc-sse-x8.c
+++ b/src/f32-vbinary/gen/vminc-sse-x8.c
@@ -20,13 +20,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const __m128 vy_min = _mm_load_ps(params->sse.min);
- const __m128 vy_max = _mm_load_ps(params->sse.max);
const __m128 vb = _mm_load1_ps(b);
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
@@ -37,11 +35,6 @@
__m128 vy0123 = _mm_min_ps(va0123, vb);
__m128 vy4567 = _mm_min_ps(va4567, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy4567 = _mm_max_ps(vy4567, vy_min);
-
- vy0123 = _mm_min_ps(vy0123, vy_max);
- vy4567 = _mm_min_ps(vy4567, vy_max);
_mm_storeu_ps(y, vy0123);
_mm_storeu_ps(y + 4, vy4567);
@@ -52,8 +45,6 @@
a += 4;
__m128 vy0123 = _mm_min_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
_mm_storeu_ps(y, vy0123);
y += 4;
}
@@ -61,8 +52,6 @@
const __m128 va0123 = _mm_loadu_ps(a);
__m128 vy0123 = _mm_min_ps(va0123, vb);
- vy0123 = _mm_max_ps(vy0123, vy_min);
- vy0123 = _mm_min_ps(vy0123, vy_max);
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy0123);
vy0123 = _mm_movehl_ps(vy0123, vy0123);
diff --git a/src/f32-vbinary/gen/vminc-wasm-x1.c b/src/f32-vbinary/gen/vminc-wasm-x1.c
index 3021406..13179ed 100644
--- a/src/f32-vbinary/gen/vminc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vminc-wasm-x1.c
@@ -19,20 +19,16 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= sizeof(float); n -= sizeof(float)) {
const float va = *a++;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
}
}
diff --git a/src/f32-vbinary/gen/vminc-wasm-x2.c b/src/f32-vbinary/gen/vminc-wasm-x2.c
index f8a99b2..f2d6744 100644
--- a/src/f32-vbinary/gen/vminc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vminc-wasm-x2.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 2 * sizeof(float); n -= 2 * sizeof(float)) {
@@ -36,11 +34,6 @@
float vy0 = __builtin_wasm_min_f32(va0, vb);
float vy1 = __builtin_wasm_min_f32(va1, vb);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -49,8 +42,6 @@
if XNN_UNLIKELY(n != 0) {
const float va = *a;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y = vy;
}
}
diff --git a/src/f32-vbinary/gen/vminc-wasm-x4.c b/src/f32-vbinary/gen/vminc-wasm-x4.c
index 5237eef..d6e3266 100644
--- a/src/f32-vbinary/gen/vminc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vminc-wasm-x4.c
@@ -19,13 +19,11 @@
const float* a,
const float* b,
float* y,
- const union xnn_f32_minmax_params params[restrict static 1])
+ const union xnn_f32_default_params params[restrict static 1])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
- const float vy_min = params->scalar.min;
- const float vy_max = params->scalar.max;
const float vb = *b;
for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
@@ -40,15 +38,6 @@
float vy2 = __builtin_wasm_min_f32(va2, vb);
float vy3 = __builtin_wasm_min_f32(va3, vb);
- vy0 = __builtin_wasm_max_f32(vy0, vy_min);
- vy1 = __builtin_wasm_max_f32(vy1, vy_min);
- vy2 = __builtin_wasm_max_f32(vy2, vy_min);
- vy3 = __builtin_wasm_max_f32(vy3, vy_min);
-
- vy0 = __builtin_wasm_min_f32(vy0, vy_max);
- vy1 = __builtin_wasm_min_f32(vy1, vy_max);
- vy2 = __builtin_wasm_min_f32(vy2, vy_max);
- vy3 = __builtin_wasm_min_f32(vy3, vy_max);
y[0] = vy0;
y[1] = vy1;
@@ -60,8 +49,6 @@
do {
const float va = *a++;
float vy = __builtin_wasm_min_f32(va, vb);
- vy = __builtin_wasm_max_f32(vy, vy_min);
- vy = __builtin_wasm_min_f32(vy, vy_max);
*y++ = vy;
n -= sizeof(float);
} while (n != 0);
diff --git a/src/f32-vbinary/gen/vmul-avx-x16.c b/src/f32-vbinary/gen/vmul-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vmul-avx-x16.c
rename to src/f32-vbinary/gen/vmul-minmax-avx-x16.c
index 8f4eee0..a51e024 100644
--- a/src/f32-vbinary/gen/vmul-avx-x16.c
+++ b/src/f32-vbinary/gen/vmul-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vmul_ukernel__avx_x16(
+void xnn_f32_vmul_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-avx-x8.c b/src/f32-vbinary/gen/vmul-minmax-avx-x8.c
similarity index 98%
rename from src/f32-vbinary/gen/vmul-avx-x8.c
rename to src/f32-vbinary/gen/vmul-minmax-avx-x8.c
index 14ffbe2..a7877c3 100644
--- a/src/f32-vbinary/gen/vmul-avx-x8.c
+++ b/src/f32-vbinary/gen/vmul-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vmul_ukernel__avx_x8(
+void xnn_f32_vmul_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-avx512f-x16.c b/src/f32-vbinary/gen/vmul-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-avx512f-x16.c
rename to src/f32-vbinary/gen/vmul-minmax-avx512f-x16.c
index 6555357..96514cd 100644
--- a/src/f32-vbinary/gen/vmul-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vmul-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__avx512f_x16(
+void xnn_f32_vmul_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-avx512f-x32.c b/src/f32-vbinary/gen/vmul-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-avx512f-x32.c
rename to src/f32-vbinary/gen/vmul-minmax-avx512f-x32.c
index 729d9d8..ef431de 100644
--- a/src/f32-vbinary/gen/vmul-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vmul-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__avx512f_x32(
+void xnn_f32_vmul_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-neon-x4.c b/src/f32-vbinary/gen/vmul-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-neon-x4.c
rename to src/f32-vbinary/gen/vmul-minmax-neon-x4.c
index 111d786..f836f03 100644
--- a/src/f32-vbinary/gen/vmul-neon-x4.c
+++ b/src/f32-vbinary/gen/vmul-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__neon_x4(
+void xnn_f32_vmul_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-neon-x8.c b/src/f32-vbinary/gen/vmul-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-neon-x8.c
rename to src/f32-vbinary/gen/vmul-minmax-neon-x8.c
index eab3025..a3d310b 100644
--- a/src/f32-vbinary/gen/vmul-neon-x8.c
+++ b/src/f32-vbinary/gen/vmul-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__neon_x8(
+void xnn_f32_vmul_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-psimd-x4.c b/src/f32-vbinary/gen/vmul-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-psimd-x4.c
rename to src/f32-vbinary/gen/vmul-minmax-psimd-x4.c
index ebfe03a..89753b4 100644
--- a/src/f32-vbinary/gen/vmul-psimd-x4.c
+++ b/src/f32-vbinary/gen/vmul-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__psimd_x4(
+void xnn_f32_vmul_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-psimd-x8.c b/src/f32-vbinary/gen/vmul-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-psimd-x8.c
rename to src/f32-vbinary/gen/vmul-minmax-psimd-x8.c
index 50e86cb..36fb4cd 100644
--- a/src/f32-vbinary/gen/vmul-psimd-x8.c
+++ b/src/f32-vbinary/gen/vmul-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__psimd_x8(
+void xnn_f32_vmul_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-scalar-x1.c b/src/f32-vbinary/gen/vmul-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vmul-scalar-x1.c
rename to src/f32-vbinary/gen/vmul-minmax-scalar-x1.c
index dd59b58..f0990f6 100644
--- a/src/f32-vbinary/gen/vmul-scalar-x1.c
+++ b/src/f32-vbinary/gen/vmul-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__scalar_x1(
+void xnn_f32_vmul_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-scalar-x2.c b/src/f32-vbinary/gen/vmul-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vmul-scalar-x2.c
rename to src/f32-vbinary/gen/vmul-minmax-scalar-x2.c
index 8c39c2f..e03a0f8 100644
--- a/src/f32-vbinary/gen/vmul-scalar-x2.c
+++ b/src/f32-vbinary/gen/vmul-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__scalar_x2(
+void xnn_f32_vmul_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-scalar-x4.c b/src/f32-vbinary/gen/vmul-minmax-scalar-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-scalar-x4.c
rename to src/f32-vbinary/gen/vmul-minmax-scalar-x4.c
index 5d849c1..dbd2db2 100644
--- a/src/f32-vbinary/gen/vmul-scalar-x4.c
+++ b/src/f32-vbinary/gen/vmul-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__scalar_x4(
+void xnn_f32_vmul_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-sse-x4.c b/src/f32-vbinary/gen/vmul-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-sse-x4.c
rename to src/f32-vbinary/gen/vmul-minmax-sse-x4.c
index 6c5b186..42f8f5f 100644
--- a/src/f32-vbinary/gen/vmul-sse-x4.c
+++ b/src/f32-vbinary/gen/vmul-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__sse_x4(
+void xnn_f32_vmul_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-sse-x8.c b/src/f32-vbinary/gen/vmul-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-sse-x8.c
rename to src/f32-vbinary/gen/vmul-minmax-sse-x8.c
index bb3930a..9631ef5 100644
--- a/src/f32-vbinary/gen/vmul-sse-x8.c
+++ b/src/f32-vbinary/gen/vmul-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__sse_x8(
+void xnn_f32_vmul_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-wasm-x1.c b/src/f32-vbinary/gen/vmul-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vmul-wasm-x1.c
rename to src/f32-vbinary/gen/vmul-minmax-wasm-x1.c
index beee3c8..d230195 100644
--- a/src/f32-vbinary/gen/vmul-wasm-x1.c
+++ b/src/f32-vbinary/gen/vmul-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__wasm_x1(
+void xnn_f32_vmul_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-wasm-x2.c b/src/f32-vbinary/gen/vmul-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vmul-wasm-x2.c
rename to src/f32-vbinary/gen/vmul-minmax-wasm-x2.c
index 6212b2f..187c9f0 100644
--- a/src/f32-vbinary/gen/vmul-wasm-x2.c
+++ b/src/f32-vbinary/gen/vmul-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__wasm_x2(
+void xnn_f32_vmul_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmul-wasm-x4.c b/src/f32-vbinary/gen/vmul-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmul-wasm-x4.c
rename to src/f32-vbinary/gen/vmul-minmax-wasm-x4.c
index 7fba937..f2d468b 100644
--- a/src/f32-vbinary/gen/vmul-wasm-x4.c
+++ b/src/f32-vbinary/gen/vmul-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmul_ukernel__wasm_x4(
+void xnn_f32_vmul_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-avx-x16.c b/src/f32-vbinary/gen/vmulc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vmulc-avx-x16.c
rename to src/f32-vbinary/gen/vmulc-minmax-avx-x16.c
index f0de274..62ba61c 100644
--- a/src/f32-vbinary/gen/vmulc-avx-x16.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vmulc_ukernel__avx_x16(
+void xnn_f32_vmulc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-avx-x8.c b/src/f32-vbinary/gen/vmulc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-avx-x8.c
rename to src/f32-vbinary/gen/vmulc-minmax-avx-x8.c
index 63aa806..736495d 100644
--- a/src/f32-vbinary/gen/vmulc-avx-x8.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vmulc_ukernel__avx_x8(
+void xnn_f32_vmulc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-avx512f-x16.c b/src/f32-vbinary/gen/vmulc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-avx512f-x16.c
rename to src/f32-vbinary/gen/vmulc-minmax-avx512f-x16.c
index b75cf0e..7eaea27 100644
--- a/src/f32-vbinary/gen/vmulc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__avx512f_x16(
+void xnn_f32_vmulc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-avx512f-x32.c b/src/f32-vbinary/gen/vmulc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-avx512f-x32.c
rename to src/f32-vbinary/gen/vmulc-minmax-avx512f-x32.c
index 328e159..51fea4d 100644
--- a/src/f32-vbinary/gen/vmulc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__avx512f_x32(
+void xnn_f32_vmulc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-neon-x4.c b/src/f32-vbinary/gen/vmulc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-neon-x4.c
rename to src/f32-vbinary/gen/vmulc-minmax-neon-x4.c
index f27dece..7782e11 100644
--- a/src/f32-vbinary/gen/vmulc-neon-x4.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__neon_x4(
+void xnn_f32_vmulc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-neon-x8.c b/src/f32-vbinary/gen/vmulc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-neon-x8.c
rename to src/f32-vbinary/gen/vmulc-minmax-neon-x8.c
index 5759595..f19dd92 100644
--- a/src/f32-vbinary/gen/vmulc-neon-x8.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__neon_x8(
+void xnn_f32_vmulc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-psimd-x4.c b/src/f32-vbinary/gen/vmulc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-psimd-x4.c
rename to src/f32-vbinary/gen/vmulc-minmax-psimd-x4.c
index 57062a4..15d5b68 100644
--- a/src/f32-vbinary/gen/vmulc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__psimd_x4(
+void xnn_f32_vmulc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-psimd-x8.c b/src/f32-vbinary/gen/vmulc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-psimd-x8.c
rename to src/f32-vbinary/gen/vmulc-minmax-psimd-x8.c
index 4929700..70412ef 100644
--- a/src/f32-vbinary/gen/vmulc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__psimd_x8(
+void xnn_f32_vmulc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-scalar-x1.c b/src/f32-vbinary/gen/vmulc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vmulc-scalar-x1.c
rename to src/f32-vbinary/gen/vmulc-minmax-scalar-x1.c
index 94859c8..005b7c28 100644
--- a/src/f32-vbinary/gen/vmulc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__scalar_x1(
+void xnn_f32_vmulc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-scalar-x2.c b/src/f32-vbinary/gen/vmulc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vmulc-scalar-x2.c
rename to src/f32-vbinary/gen/vmulc-minmax-scalar-x2.c
index 1fff64b..dd9f1ea 100644
--- a/src/f32-vbinary/gen/vmulc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__scalar_x2(
+void xnn_f32_vmulc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-scalar-x4.c b/src/f32-vbinary/gen/vmulc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vmulc-scalar-x4.c
rename to src/f32-vbinary/gen/vmulc-minmax-scalar-x4.c
index e52e472..c39a7bd 100644
--- a/src/f32-vbinary/gen/vmulc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__scalar_x4(
+void xnn_f32_vmulc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-sse-x4.c b/src/f32-vbinary/gen/vmulc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-sse-x4.c
rename to src/f32-vbinary/gen/vmulc-minmax-sse-x4.c
index 6bc225c..4887488 100644
--- a/src/f32-vbinary/gen/vmulc-sse-x4.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__sse_x4(
+void xnn_f32_vmulc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-sse-x8.c b/src/f32-vbinary/gen/vmulc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-sse-x8.c
rename to src/f32-vbinary/gen/vmulc-minmax-sse-x8.c
index 40ba8d2..5be725a 100644
--- a/src/f32-vbinary/gen/vmulc-sse-x8.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__sse_x8(
+void xnn_f32_vmulc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-wasm-x1.c b/src/f32-vbinary/gen/vmulc-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vmulc-wasm-x1.c
rename to src/f32-vbinary/gen/vmulc-minmax-wasm-x1.c
index cfe73b1..20b61ff 100644
--- a/src/f32-vbinary/gen/vmulc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__wasm_x1(
+void xnn_f32_vmulc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-wasm-x2.c b/src/f32-vbinary/gen/vmulc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vmulc-wasm-x2.c
rename to src/f32-vbinary/gen/vmulc-minmax-wasm-x2.c
index 5baab70..533101f 100644
--- a/src/f32-vbinary/gen/vmulc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__wasm_x2(
+void xnn_f32_vmulc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vmulc-wasm-x4.c b/src/f32-vbinary/gen/vmulc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vmulc-wasm-x4.c
rename to src/f32-vbinary/gen/vmulc-minmax-wasm-x4.c
index 3c3fe0a..bcdbee9 100644
--- a/src/f32-vbinary/gen/vmulc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vmulc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vmulc_ukernel__wasm_x4(
+void xnn_f32_vmulc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-avx-x16.c b/src/f32-vbinary/gen/vrdivc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vrdivc-avx-x16.c
rename to src/f32-vbinary/gen/vrdivc-minmax-avx-x16.c
index ac194b5..653eb1a 100644
--- a/src/f32-vbinary/gen/vrdivc-avx-x16.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vrdivc_ukernel__avx_x16(
+void xnn_f32_vrdivc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-avx-x8.c b/src/f32-vbinary/gen/vrdivc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-avx-x8.c
rename to src/f32-vbinary/gen/vrdivc-minmax-avx-x8.c
index 6259bce..ed2350a 100644
--- a/src/f32-vbinary/gen/vrdivc-avx-x8.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vrdivc_ukernel__avx_x8(
+void xnn_f32_vrdivc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-avx512f-x16.c b/src/f32-vbinary/gen/vrdivc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-avx512f-x16.c
rename to src/f32-vbinary/gen/vrdivc-minmax-avx512f-x16.c
index 353d011..04f9506 100644
--- a/src/f32-vbinary/gen/vrdivc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__avx512f_x16(
+void xnn_f32_vrdivc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-avx512f-x32.c b/src/f32-vbinary/gen/vrdivc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-avx512f-x32.c
rename to src/f32-vbinary/gen/vrdivc-minmax-avx512f-x32.c
index 3d37404..27a66f0 100644
--- a/src/f32-vbinary/gen/vrdivc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__avx512f_x32(
+void xnn_f32_vrdivc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-neon-x4.c b/src/f32-vbinary/gen/vrdivc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-neon-x4.c
rename to src/f32-vbinary/gen/vrdivc-minmax-neon-x4.c
index 529ab86..089fc74 100644
--- a/src/f32-vbinary/gen/vrdivc-neon-x4.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__neon_x4(
+void xnn_f32_vrdivc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-neon-x8.c b/src/f32-vbinary/gen/vrdivc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-neon-x8.c
rename to src/f32-vbinary/gen/vrdivc-minmax-neon-x8.c
index 2e5bccb..341835f 100644
--- a/src/f32-vbinary/gen/vrdivc-neon-x8.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__neon_x8(
+void xnn_f32_vrdivc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-psimd-x4.c b/src/f32-vbinary/gen/vrdivc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-psimd-x4.c
rename to src/f32-vbinary/gen/vrdivc-minmax-psimd-x4.c
index cf5f1a2..0db3288 100644
--- a/src/f32-vbinary/gen/vrdivc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__psimd_x4(
+void xnn_f32_vrdivc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-psimd-x8.c b/src/f32-vbinary/gen/vrdivc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-psimd-x8.c
rename to src/f32-vbinary/gen/vrdivc-minmax-psimd-x8.c
index b5ec17a..b03c4f5 100644
--- a/src/f32-vbinary/gen/vrdivc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__psimd_x8(
+void xnn_f32_vrdivc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-scalar-x1.c b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vrdivc-scalar-x1.c
rename to src/f32-vbinary/gen/vrdivc-minmax-scalar-x1.c
index 7ce55b5..7de486e 100644
--- a/src/f32-vbinary/gen/vrdivc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__scalar_x1(
+void xnn_f32_vrdivc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-scalar-x2.c b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vrdivc-scalar-x2.c
rename to src/f32-vbinary/gen/vrdivc-minmax-scalar-x2.c
index c0bbb10..26856f0 100644
--- a/src/f32-vbinary/gen/vrdivc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__scalar_x2(
+void xnn_f32_vrdivc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-scalar-x4.c b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vrdivc-scalar-x4.c
rename to src/f32-vbinary/gen/vrdivc-minmax-scalar-x4.c
index 89dcfa8..26a8bd9 100644
--- a/src/f32-vbinary/gen/vrdivc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__scalar_x4(
+void xnn_f32_vrdivc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-sse-x4.c b/src/f32-vbinary/gen/vrdivc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-sse-x4.c
rename to src/f32-vbinary/gen/vrdivc-minmax-sse-x4.c
index 0a036ca..36865b9 100644
--- a/src/f32-vbinary/gen/vrdivc-sse-x4.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__sse_x4(
+void xnn_f32_vrdivc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-sse-x8.c b/src/f32-vbinary/gen/vrdivc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-sse-x8.c
rename to src/f32-vbinary/gen/vrdivc-minmax-sse-x8.c
index 17f2688..793729d 100644
--- a/src/f32-vbinary/gen/vrdivc-sse-x8.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__sse_x8(
+void xnn_f32_vrdivc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-wasm-x1.c b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vrdivc-wasm-x1.c
rename to src/f32-vbinary/gen/vrdivc-minmax-wasm-x1.c
index f66479b..2013103 100644
--- a/src/f32-vbinary/gen/vrdivc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__wasm_x1(
+void xnn_f32_vrdivc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-wasm-x2.c b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vrdivc-wasm-x2.c
rename to src/f32-vbinary/gen/vrdivc-minmax-wasm-x2.c
index ec02f6d..5e47615 100644
--- a/src/f32-vbinary/gen/vrdivc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__wasm_x2(
+void xnn_f32_vrdivc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrdivc-wasm-x4.c b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrdivc-wasm-x4.c
rename to src/f32-vbinary/gen/vrdivc-minmax-wasm-x4.c
index 7318952..4e8bd1a 100644
--- a/src/f32-vbinary/gen/vrdivc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vrdivc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrdivc_ukernel__wasm_x4(
+void xnn_f32_vrdivc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-avx-x16.c b/src/f32-vbinary/gen/vrsubc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vrsubc-avx-x16.c
rename to src/f32-vbinary/gen/vrsubc-minmax-avx-x16.c
index a057d71..1bf1fe4 100644
--- a/src/f32-vbinary/gen/vrsubc-avx-x16.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vrsubc_ukernel__avx_x16(
+void xnn_f32_vrsubc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-avx-x8.c b/src/f32-vbinary/gen/vrsubc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-avx-x8.c
rename to src/f32-vbinary/gen/vrsubc-minmax-avx-x8.c
index 4fdc243..d1b7bbd 100644
--- a/src/f32-vbinary/gen/vrsubc-avx-x8.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vrsubc_ukernel__avx_x8(
+void xnn_f32_vrsubc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-avx512f-x16.c b/src/f32-vbinary/gen/vrsubc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-avx512f-x16.c
rename to src/f32-vbinary/gen/vrsubc-minmax-avx512f-x16.c
index 2098cb0..c158285 100644
--- a/src/f32-vbinary/gen/vrsubc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__avx512f_x16(
+void xnn_f32_vrsubc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-avx512f-x32.c b/src/f32-vbinary/gen/vrsubc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-avx512f-x32.c
rename to src/f32-vbinary/gen/vrsubc-minmax-avx512f-x32.c
index 7f57f2c..110d4b8 100644
--- a/src/f32-vbinary/gen/vrsubc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__avx512f_x32(
+void xnn_f32_vrsubc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-neon-x4.c b/src/f32-vbinary/gen/vrsubc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-neon-x4.c
rename to src/f32-vbinary/gen/vrsubc-minmax-neon-x4.c
index 6864d32..517516a 100644
--- a/src/f32-vbinary/gen/vrsubc-neon-x4.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__neon_x4(
+void xnn_f32_vrsubc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-neon-x8.c b/src/f32-vbinary/gen/vrsubc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-neon-x8.c
rename to src/f32-vbinary/gen/vrsubc-minmax-neon-x8.c
index 6f37f74..fdd6b25 100644
--- a/src/f32-vbinary/gen/vrsubc-neon-x8.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__neon_x8(
+void xnn_f32_vrsubc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-psimd-x4.c b/src/f32-vbinary/gen/vrsubc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-psimd-x4.c
rename to src/f32-vbinary/gen/vrsubc-minmax-psimd-x4.c
index a2c660b..01e3c41 100644
--- a/src/f32-vbinary/gen/vrsubc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__psimd_x4(
+void xnn_f32_vrsubc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-psimd-x8.c b/src/f32-vbinary/gen/vrsubc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-psimd-x8.c
rename to src/f32-vbinary/gen/vrsubc-minmax-psimd-x8.c
index aeb68e6..8aa2363 100644
--- a/src/f32-vbinary/gen/vrsubc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__psimd_x8(
+void xnn_f32_vrsubc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-scalar-x1.c b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vrsubc-scalar-x1.c
rename to src/f32-vbinary/gen/vrsubc-minmax-scalar-x1.c
index fab1e7c..785943b 100644
--- a/src/f32-vbinary/gen/vrsubc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__scalar_x1(
+void xnn_f32_vrsubc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-scalar-x2.c b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vrsubc-scalar-x2.c
rename to src/f32-vbinary/gen/vrsubc-minmax-scalar-x2.c
index f82099e..5e975f5 100644
--- a/src/f32-vbinary/gen/vrsubc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__scalar_x2(
+void xnn_f32_vrsubc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-scalar-x4.c b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vrsubc-scalar-x4.c
rename to src/f32-vbinary/gen/vrsubc-minmax-scalar-x4.c
index 3a0c5a9..b029113 100644
--- a/src/f32-vbinary/gen/vrsubc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__scalar_x4(
+void xnn_f32_vrsubc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-sse-x4.c b/src/f32-vbinary/gen/vrsubc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-sse-x4.c
rename to src/f32-vbinary/gen/vrsubc-minmax-sse-x4.c
index ad343ae..e8e944b 100644
--- a/src/f32-vbinary/gen/vrsubc-sse-x4.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__sse_x4(
+void xnn_f32_vrsubc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-sse-x8.c b/src/f32-vbinary/gen/vrsubc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-sse-x8.c
rename to src/f32-vbinary/gen/vrsubc-minmax-sse-x8.c
index 75e9aca..75f2a44 100644
--- a/src/f32-vbinary/gen/vrsubc-sse-x8.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__sse_x8(
+void xnn_f32_vrsubc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-wasm-x1.c b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vrsubc-wasm-x1.c
rename to src/f32-vbinary/gen/vrsubc-minmax-wasm-x1.c
index 5849e4d..d844d6d 100644
--- a/src/f32-vbinary/gen/vrsubc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__wasm_x1(
+void xnn_f32_vrsubc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-wasm-x2.c b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vrsubc-wasm-x2.c
rename to src/f32-vbinary/gen/vrsubc-minmax-wasm-x2.c
index 7f94c23..8931015 100644
--- a/src/f32-vbinary/gen/vrsubc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__wasm_x2(
+void xnn_f32_vrsubc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vrsubc-wasm-x4.c b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vrsubc-wasm-x4.c
rename to src/f32-vbinary/gen/vrsubc-minmax-wasm-x4.c
index bec931c..1735af2 100644
--- a/src/f32-vbinary/gen/vrsubc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vrsubc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vrsubc_ukernel__wasm_x4(
+void xnn_f32_vrsubc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-avx-x16.c b/src/f32-vbinary/gen/vsub-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vsub-avx-x16.c
rename to src/f32-vbinary/gen/vsub-minmax-avx-x16.c
index 9beabcd..386aca3 100644
--- a/src/f32-vbinary/gen/vsub-avx-x16.c
+++ b/src/f32-vbinary/gen/vsub-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vsub_ukernel__avx_x16(
+void xnn_f32_vsub_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-avx-x8.c b/src/f32-vbinary/gen/vsub-minmax-avx-x8.c
similarity index 98%
rename from src/f32-vbinary/gen/vsub-avx-x8.c
rename to src/f32-vbinary/gen/vsub-minmax-avx-x8.c
index bdaf5cb..cb063fe 100644
--- a/src/f32-vbinary/gen/vsub-avx-x8.c
+++ b/src/f32-vbinary/gen/vsub-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vsub_ukernel__avx_x8(
+void xnn_f32_vsub_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-avx512f-x16.c b/src/f32-vbinary/gen/vsub-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-avx512f-x16.c
rename to src/f32-vbinary/gen/vsub-minmax-avx512f-x16.c
index be9ba7b..eb639dd 100644
--- a/src/f32-vbinary/gen/vsub-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vsub-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__avx512f_x16(
+void xnn_f32_vsub_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-avx512f-x32.c b/src/f32-vbinary/gen/vsub-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-avx512f-x32.c
rename to src/f32-vbinary/gen/vsub-minmax-avx512f-x32.c
index 11879c2..40515e5 100644
--- a/src/f32-vbinary/gen/vsub-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vsub-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__avx512f_x32(
+void xnn_f32_vsub_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-neon-x4.c b/src/f32-vbinary/gen/vsub-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-neon-x4.c
rename to src/f32-vbinary/gen/vsub-minmax-neon-x4.c
index c96b4a0..8a55ada 100644
--- a/src/f32-vbinary/gen/vsub-neon-x4.c
+++ b/src/f32-vbinary/gen/vsub-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__neon_x4(
+void xnn_f32_vsub_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-neon-x8.c b/src/f32-vbinary/gen/vsub-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-neon-x8.c
rename to src/f32-vbinary/gen/vsub-minmax-neon-x8.c
index 205bc84..04ff548 100644
--- a/src/f32-vbinary/gen/vsub-neon-x8.c
+++ b/src/f32-vbinary/gen/vsub-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__neon_x8(
+void xnn_f32_vsub_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-psimd-x4.c b/src/f32-vbinary/gen/vsub-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-psimd-x4.c
rename to src/f32-vbinary/gen/vsub-minmax-psimd-x4.c
index 35529f6..670457b 100644
--- a/src/f32-vbinary/gen/vsub-psimd-x4.c
+++ b/src/f32-vbinary/gen/vsub-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__psimd_x4(
+void xnn_f32_vsub_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-psimd-x8.c b/src/f32-vbinary/gen/vsub-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-psimd-x8.c
rename to src/f32-vbinary/gen/vsub-minmax-psimd-x8.c
index 0b50884..505e639 100644
--- a/src/f32-vbinary/gen/vsub-psimd-x8.c
+++ b/src/f32-vbinary/gen/vsub-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__psimd_x8(
+void xnn_f32_vsub_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-scalar-x1.c b/src/f32-vbinary/gen/vsub-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vsub-scalar-x1.c
rename to src/f32-vbinary/gen/vsub-minmax-scalar-x1.c
index fcf9db7..d39e066 100644
--- a/src/f32-vbinary/gen/vsub-scalar-x1.c
+++ b/src/f32-vbinary/gen/vsub-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__scalar_x1(
+void xnn_f32_vsub_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-scalar-x2.c b/src/f32-vbinary/gen/vsub-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vsub-scalar-x2.c
rename to src/f32-vbinary/gen/vsub-minmax-scalar-x2.c
index ff80f2c..52f9de1 100644
--- a/src/f32-vbinary/gen/vsub-scalar-x2.c
+++ b/src/f32-vbinary/gen/vsub-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__scalar_x2(
+void xnn_f32_vsub_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-scalar-x4.c b/src/f32-vbinary/gen/vsub-minmax-scalar-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-scalar-x4.c
rename to src/f32-vbinary/gen/vsub-minmax-scalar-x4.c
index af50403..385fe9c 100644
--- a/src/f32-vbinary/gen/vsub-scalar-x4.c
+++ b/src/f32-vbinary/gen/vsub-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__scalar_x4(
+void xnn_f32_vsub_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-sse-x4.c b/src/f32-vbinary/gen/vsub-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-sse-x4.c
rename to src/f32-vbinary/gen/vsub-minmax-sse-x4.c
index 5a1a904..29b54c0 100644
--- a/src/f32-vbinary/gen/vsub-sse-x4.c
+++ b/src/f32-vbinary/gen/vsub-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__sse_x4(
+void xnn_f32_vsub_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-sse-x8.c b/src/f32-vbinary/gen/vsub-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-sse-x8.c
rename to src/f32-vbinary/gen/vsub-minmax-sse-x8.c
index 35efaa5..24d2acc 100644
--- a/src/f32-vbinary/gen/vsub-sse-x8.c
+++ b/src/f32-vbinary/gen/vsub-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__sse_x8(
+void xnn_f32_vsub_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-wasm-x1.c b/src/f32-vbinary/gen/vsub-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vsub-wasm-x1.c
rename to src/f32-vbinary/gen/vsub-minmax-wasm-x1.c
index e8589d7..13b0c7d 100644
--- a/src/f32-vbinary/gen/vsub-wasm-x1.c
+++ b/src/f32-vbinary/gen/vsub-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__wasm_x1(
+void xnn_f32_vsub_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-wasm-x2.c b/src/f32-vbinary/gen/vsub-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vsub-wasm-x2.c
rename to src/f32-vbinary/gen/vsub-minmax-wasm-x2.c
index 8c0a332..c76fdee 100644
--- a/src/f32-vbinary/gen/vsub-wasm-x2.c
+++ b/src/f32-vbinary/gen/vsub-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__wasm_x2(
+void xnn_f32_vsub_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsub-wasm-x4.c b/src/f32-vbinary/gen/vsub-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsub-wasm-x4.c
rename to src/f32-vbinary/gen/vsub-minmax-wasm-x4.c
index 522170d..616cfc5 100644
--- a/src/f32-vbinary/gen/vsub-wasm-x4.c
+++ b/src/f32-vbinary/gen/vsub-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsub_ukernel__wasm_x4(
+void xnn_f32_vsub_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-avx-x16.c b/src/f32-vbinary/gen/vsubc-minmax-avx-x16.c
similarity index 98%
rename from src/f32-vbinary/gen/vsubc-avx-x16.c
rename to src/f32-vbinary/gen/vsubc-minmax-avx-x16.c
index 4dc04b6..6483a19 100644
--- a/src/f32-vbinary/gen/vsubc-avx-x16.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-avx-x16.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vsubc_ukernel__avx_x16(
+void xnn_f32_vsubc_minmax_ukernel__avx_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-avx-x8.c b/src/f32-vbinary/gen/vsubc-minmax-avx-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-avx-x8.c
rename to src/f32-vbinary/gen/vsubc-minmax-avx-x8.c
index e9c3bbd..192096f 100644
--- a/src/f32-vbinary/gen/vsubc-avx-x8.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-avx-x8.c
@@ -17,7 +17,7 @@
static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
-void xnn_f32_vsubc_ukernel__avx_x8(
+void xnn_f32_vsubc_minmax_ukernel__avx_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-avx512f-x16.c b/src/f32-vbinary/gen/vsubc-minmax-avx512f-x16.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-avx512f-x16.c
rename to src/f32-vbinary/gen/vsubc-minmax-avx512f-x16.c
index 37d1a6f..71a9fcf 100644
--- a/src/f32-vbinary/gen/vsubc-avx512f-x16.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-avx512f-x16.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__avx512f_x16(
+void xnn_f32_vsubc_minmax_ukernel__avx512f_x16(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-avx512f-x32.c b/src/f32-vbinary/gen/vsubc-minmax-avx512f-x32.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-avx512f-x32.c
rename to src/f32-vbinary/gen/vsubc-minmax-avx512f-x32.c
index 3fcb862..bf4354e 100644
--- a/src/f32-vbinary/gen/vsubc-avx512f-x32.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-avx512f-x32.c
@@ -16,7 +16,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__avx512f_x32(
+void xnn_f32_vsubc_minmax_ukernel__avx512f_x32(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-neon-x4.c b/src/f32-vbinary/gen/vsubc-minmax-neon-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-neon-x4.c
rename to src/f32-vbinary/gen/vsubc-minmax-neon-x4.c
index 531324d..0ffcc23 100644
--- a/src/f32-vbinary/gen/vsubc-neon-x4.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-neon-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__neon_x4(
+void xnn_f32_vsubc_minmax_ukernel__neon_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-neon-x8.c b/src/f32-vbinary/gen/vsubc-minmax-neon-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-neon-x8.c
rename to src/f32-vbinary/gen/vsubc-minmax-neon-x8.c
index 9905f56..09d9f0f 100644
--- a/src/f32-vbinary/gen/vsubc-neon-x8.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-neon-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__neon_x8(
+void xnn_f32_vsubc_minmax_ukernel__neon_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-psimd-x4.c b/src/f32-vbinary/gen/vsubc-minmax-psimd-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-psimd-x4.c
rename to src/f32-vbinary/gen/vsubc-minmax-psimd-x4.c
index be5908e..feb9c88 100644
--- a/src/f32-vbinary/gen/vsubc-psimd-x4.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-psimd-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__psimd_x4(
+void xnn_f32_vsubc_minmax_ukernel__psimd_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-psimd-x8.c b/src/f32-vbinary/gen/vsubc-minmax-psimd-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-psimd-x8.c
rename to src/f32-vbinary/gen/vsubc-minmax-psimd-x8.c
index 1939191..3206be9 100644
--- a/src/f32-vbinary/gen/vsubc-psimd-x8.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-psimd-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__psimd_x8(
+void xnn_f32_vsubc_minmax_ukernel__psimd_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-scalar-x1.c b/src/f32-vbinary/gen/vsubc-minmax-scalar-x1.c
similarity index 94%
rename from src/f32-vbinary/gen/vsubc-scalar-x1.c
rename to src/f32-vbinary/gen/vsubc-minmax-scalar-x1.c
index 9da9158..a476df5 100644
--- a/src/f32-vbinary/gen/vsubc-scalar-x1.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-scalar-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__scalar_x1(
+void xnn_f32_vsubc_minmax_ukernel__scalar_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-scalar-x2.c b/src/f32-vbinary/gen/vsubc-minmax-scalar-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vsubc-scalar-x2.c
rename to src/f32-vbinary/gen/vsubc-minmax-scalar-x2.c
index 16d89cc..8efb9b1 100644
--- a/src/f32-vbinary/gen/vsubc-scalar-x2.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-scalar-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__scalar_x2(
+void xnn_f32_vsubc_minmax_ukernel__scalar_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-scalar-x4.c b/src/f32-vbinary/gen/vsubc-minmax-scalar-x4.c
similarity index 96%
rename from src/f32-vbinary/gen/vsubc-scalar-x4.c
rename to src/f32-vbinary/gen/vsubc-minmax-scalar-x4.c
index 4fd21d9..9a61698 100644
--- a/src/f32-vbinary/gen/vsubc-scalar-x4.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-scalar-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__scalar_x4(
+void xnn_f32_vsubc_minmax_ukernel__scalar_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-sse-x4.c b/src/f32-vbinary/gen/vsubc-minmax-sse-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-sse-x4.c
rename to src/f32-vbinary/gen/vsubc-minmax-sse-x4.c
index 62e6f0c..09c5b3a 100644
--- a/src/f32-vbinary/gen/vsubc-sse-x4.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-sse-x4.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__sse_x4(
+void xnn_f32_vsubc_minmax_ukernel__sse_x4(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-sse-x8.c b/src/f32-vbinary/gen/vsubc-minmax-sse-x8.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-sse-x8.c
rename to src/f32-vbinary/gen/vsubc-minmax-sse-x8.c
index aef2d08..bc8631d 100644
--- a/src/f32-vbinary/gen/vsubc-sse-x8.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-sse-x8.c
@@ -15,7 +15,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__sse_x8(
+void xnn_f32_vsubc_minmax_ukernel__sse_x8(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-wasm-x1.c b/src/f32-vbinary/gen/vsubc-minmax-wasm-x1.c
similarity index 95%
rename from src/f32-vbinary/gen/vsubc-wasm-x1.c
rename to src/f32-vbinary/gen/vsubc-minmax-wasm-x1.c
index ed9f69a..ed667c4 100644
--- a/src/f32-vbinary/gen/vsubc-wasm-x1.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-wasm-x1.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__wasm_x1(
+void xnn_f32_vsubc_minmax_ukernel__wasm_x1(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-wasm-x2.c b/src/f32-vbinary/gen/vsubc-minmax-wasm-x2.c
similarity index 96%
rename from src/f32-vbinary/gen/vsubc-wasm-x2.c
rename to src/f32-vbinary/gen/vsubc-minmax-wasm-x2.c
index 3ccd056..effabcf 100644
--- a/src/f32-vbinary/gen/vsubc-wasm-x2.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-wasm-x2.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__wasm_x2(
+void xnn_f32_vsubc_minmax_ukernel__wasm_x2(
size_t n,
const float* a,
const float* b,
diff --git a/src/f32-vbinary/gen/vsubc-wasm-x4.c b/src/f32-vbinary/gen/vsubc-minmax-wasm-x4.c
similarity index 97%
rename from src/f32-vbinary/gen/vsubc-wasm-x4.c
rename to src/f32-vbinary/gen/vsubc-minmax-wasm-x4.c
index 1ef8d16..8787a94 100644
--- a/src/f32-vbinary/gen/vsubc-wasm-x4.c
+++ b/src/f32-vbinary/gen/vsubc-minmax-wasm-x4.c
@@ -14,7 +14,7 @@
#include <xnnpack/vbinary.h>
-void xnn_f32_vsubc_ukernel__wasm_x4(
+void xnn_f32_vsubc_minmax_ukernel__wasm_x4(
size_t n,
const float* a,
const float* b,