QS8 Neon GEMM microkernel with 8 bit multiply and vpadal to accumulate
PiperOrigin-RevId: 353315852
diff --git a/src/qs8-gemm/gen/2x8c2-minmax-neon-mull-padal-dup.c b/src/qs8-gemm/gen/2x8c2-minmax-neon-mull-padal-dup.c
new file mode 100644
index 0000000..38c3506
--- /dev/null
+++ b/src/qs8-gemm/gen/2x8c2-minmax-neon-mull-padal-dup.c
@@ -0,0 +1,235 @@
+// Auto-generated file. Do not edit!
+// Template: src/qs8-gemm/c2-neon-mull-padal-dup.c.in
+// Generator: tools/xngen
+//
+// Copyright 2021 Google LLC
+//
+// This source code is licensed under the BSD-style license found in the
+// LICENSE file in the root directory of this source tree.
+
+#include <assert.h>
+
+#include <arm_neon.h>
+
+#include <xnnpack/common.h>
+#include <xnnpack/gemm.h>
+
+
+void xnn_qs8_gemm_minmax_ukernel_2x8c2__neon_mull_padal_dup(
+ size_t mr,
+ size_t nc,
+ size_t kc,
+ const int8_t* restrict a,
+ size_t a_stride,
+ const void* restrict w,
+ int8_t* restrict c,
+ size_t cm_stride,
+ size_t cn_stride,
+ const union xnn_qs8_gemm_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_DISABLE_TSAN
+{
+ assert(mr != 0);
+ assert(mr <= 2);
+ assert(nc != 0);
+ assert(kc != 0);
+ assert(kc % sizeof(int8_t) == 0);
+ assert(a != NULL);
+ assert(w != NULL);
+ assert(c != NULL);
+
+ const int8_t* a0 = a;
+ int8_t* c0 = c;
+ const int8_t* a1 = (const int8_t*) ((uintptr_t) a0 + a_stride);
+ int8_t* c1 = (int8_t*) ((uintptr_t) c0 + cm_stride);
+ if XNN_UNPREDICTABLE(mr != 2) {
+ a1 = a0;
+ c1 = c0;
+ }
+
+ do {
+ int32x4_t vacc0x0123 = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t));
+ int32x4_t vacc0x4567 = vld1q_s32(w); w = (const void*) ((uintptr_t) w + 4 * sizeof(int32_t));
+ int32x4_t vacc1x0123 = vacc0x0123;
+ int32x4_t vacc1x4567 = vacc0x4567;
+
+ size_t k = kc;
+ while (k >= 8 * sizeof(int8_t)) {
+ const int8x8_t va0 = vld1_s8(a0); a0 += 8;
+ const int8x8_t va1 = vld1_s8(a1); a1 += 8;
+
+ const int8x8_t vb0123c0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 0)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c0);
+ const int16x8_t vprod1x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 0)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c0);
+ const int8x8_t vb4567c0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 0)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c0);
+ const int16x8_t vprod1x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 0)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c0);
+ const int8x8_t vb0123c1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c1);
+ const int16x8_t vprod1x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 1)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c1);
+ const int8x8_t vb4567c1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c1);
+ const int16x8_t vprod1x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 1)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c1);
+ const int8x8_t vb0123c2 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c2);
+ const int16x8_t vprod1x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 2)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c2);
+ const int8x8_t vb4567c2 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c2);
+ const int16x8_t vprod1x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 2)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c2);
+ const int8x8_t vb0123c3 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 3)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c3);
+ const int16x8_t vprod1x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 3)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c3);
+ const int8x8_t vb4567c3 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 3)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c3);
+ const int16x8_t vprod1x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 3)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c3);
+
+ k -= 8 * sizeof(int8_t);
+ }
+ if XNN_UNLIKELY(k != 0) {
+ const int8x8_t va0 = vld1_s8(a0); a0 = (const int8_t*) ((uintptr_t) a0 + k);
+ const int8x8_t va1 = vld1_s8(a1); a1 = (const int8_t*) ((uintptr_t) a1 + k);
+
+ const int8x8_t vb0123c0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+ const int8x8_t vb4567c0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 0)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c0);
+ const int16x8_t vprod0x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 0)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c0);
+ const int16x8_t vprod1x0123c0 = vmull_s8(vb0123c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 0)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c0);
+ const int16x8_t vprod1x4567c0 = vmull_s8(vb4567c0, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 0)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c0);
+
+ if (k > 2 * sizeof(int8_t)) {
+ const int8x8_t vb0123c1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+ const int8x8_t vb4567c1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c1);
+ const int16x8_t vprod0x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 1)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c1);
+ const int16x8_t vprod1x0123c1 = vmull_s8(vb0123c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 1)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c1);
+ const int16x8_t vprod1x4567c1 = vmull_s8(vb4567c1, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 1)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c1);
+
+ if (k > 4 * sizeof(int8_t)) {
+ const int8x8_t vb0123c2 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+ const int8x8_t vb4567c2 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c2);
+ const int16x8_t vprod0x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 2)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c2);
+ const int16x8_t vprod1x0123c2 = vmull_s8(vb0123c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 2)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c2);
+ const int16x8_t vprod1x4567c2 = vmull_s8(vb4567c2, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 2)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c2);
+
+ if (k > 6 * sizeof(int8_t)) {
+ const int8x8_t vb0123c3 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+ const int8x8_t vb4567c3 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof(int8_t));
+
+ const int16x8_t vprod0x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 3)));
+ vacc0x0123 = vpadalq_s16(vacc0x0123, vprod0x0123c3);
+ const int16x8_t vprod0x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va0), 3)));
+ vacc0x4567 = vpadalq_s16(vacc0x4567, vprod0x4567c3);
+ const int16x8_t vprod1x0123c3 = vmull_s8(vb0123c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 3)));
+ vacc1x0123 = vpadalq_s16(vacc1x0123, vprod1x0123c3);
+ const int16x8_t vprod1x4567c3 = vmull_s8(vb4567c3, vreinterpret_s8_s16(vdup_lane_s16(vreinterpret_s16_s8(va1), 3)));
+ vacc1x4567 = vpadalq_s16(vacc1x4567, vprod1x4567c3);
+ }
+ }
+ }
+ }
+ const int32x4_t vmultiplier = vld1q_dup_s32(¶ms->neon.multiplier);
+ vacc0x0123 = vqrdmulhq_s32(vacc0x0123, vmultiplier);
+ vacc0x4567 = vqrdmulhq_s32(vacc0x4567, vmultiplier);
+ vacc1x0123 = vqrdmulhq_s32(vacc1x0123, vmultiplier);
+ vacc1x4567 = vqrdmulhq_s32(vacc1x4567, vmultiplier);
+
+ const int32x4_t vright_shift = vld1q_dup_s32(¶ms->neon.right_shift);
+ const int32x4_t vzero_shift_mask = vreinterpretq_s32_u32(vceqq_s32(vright_shift, vmovq_n_s32(0)));
+ vacc0x0123 = vsraq_n_s32(vacc0x0123, vbicq_s32(vacc0x0123, vzero_shift_mask), 31);
+ vacc0x4567 = vsraq_n_s32(vacc0x4567, vbicq_s32(vacc0x4567, vzero_shift_mask), 31);
+ vacc1x0123 = vsraq_n_s32(vacc1x0123, vbicq_s32(vacc1x0123, vzero_shift_mask), 31);
+ vacc1x4567 = vsraq_n_s32(vacc1x4567, vbicq_s32(vacc1x4567, vzero_shift_mask), 31);
+
+ vacc0x0123 = vrshlq_s32(vacc0x0123, vright_shift);
+ vacc0x4567 = vrshlq_s32(vacc0x4567, vright_shift);
+ vacc1x0123 = vrshlq_s32(vacc1x0123, vright_shift);
+ vacc1x4567 = vrshlq_s32(vacc1x4567, vright_shift);
+
+ const int16x8_t voutput_zero_point = vld1q_dup_s16(¶ms->neon.output_zero_point);
+#if XNN_ARCH_ARM64
+ const int16x8_t vacc0x01234567 = vqaddq_s16(vqmovn_high_s32(vqmovn_s32(vacc0x0123), vacc0x4567), voutput_zero_point);
+ const int16x8_t vacc1x01234567 = vqaddq_s16(vqmovn_high_s32(vqmovn_s32(vacc1x0123), vacc1x4567), voutput_zero_point);
+
+ int8x16_t vout0x01234567_1x01234567 = vqmovn_high_s16(vqmovn_s16(vacc0x01234567), vacc1x01234567);
+#else
+ const int16x8_t vacc0x01234567 = vqaddq_s16(vcombine_s16(vqmovn_s32(vacc0x0123), vqmovn_s32(vacc0x4567)), voutput_zero_point);
+ const int16x8_t vacc1x01234567 = vqaddq_s16(vcombine_s16(vqmovn_s32(vacc1x0123), vqmovn_s32(vacc1x4567)), voutput_zero_point);
+
+ int8x16_t vout0x01234567_1x01234567 = vcombine_s8(vqmovn_s16(vacc0x01234567), vqmovn_s16(vacc1x01234567));
+#endif
+ const int8x16_t voutput_min = vld1q_dup_s8(¶ms->neon.output_min);
+ const int8x16_t voutput_max = vld1q_dup_s8(¶ms->neon.output_max);
+
+ vout0x01234567_1x01234567 = vmaxq_s8(vout0x01234567_1x01234567, voutput_min);
+
+ vout0x01234567_1x01234567 = vminq_s8(vout0x01234567_1x01234567, voutput_max);
+
+ if (nc >= 8) {
+ vst1_s8(c0 + 0, vget_low_s8(vout0x01234567_1x01234567));
+ vst1_s8(c1 + 0, vget_high_s8(vout0x01234567_1x01234567));
+
+ c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
+ c1 = (int8_t*) ((uintptr_t) c1 + cn_stride);
+
+ a0 = (const int8_t*) ((uintptr_t) a0 - kc);
+ a1 = (const int8_t*) ((uintptr_t) a1 - kc);
+
+ nc -= 8;
+ } else {
+ if (nc & 4) {
+ vst1q_lane_u32(__builtin_assume_aligned(c0, 1), vreinterpretq_u32_s8(vout0x01234567_1x01234567), 0); c0 += 4;
+ vst1q_lane_u32(__builtin_assume_aligned(c1, 1), vreinterpretq_u32_s8(vout0x01234567_1x01234567), 2); c1 += 4;
+ vout0x01234567_1x01234567 = vextq_s8(vout0x01234567_1x01234567, vout0x01234567_1x01234567, 4);
+ }
+ if (nc & 2) {
+ vst1q_lane_u16(__builtin_assume_aligned(c0, 1), vreinterpretq_u16_s8(vout0x01234567_1x01234567), 0); c0 += 2;
+ vst1q_lane_u16(__builtin_assume_aligned(c1, 1), vreinterpretq_u16_s8(vout0x01234567_1x01234567), 4); c1 += 2;
+ vout0x01234567_1x01234567 = vextq_s8(vout0x01234567_1x01234567, vout0x01234567_1x01234567, 2);
+ }
+ if (nc & 1) {
+ vst1q_lane_s8(c0, vout0x01234567_1x01234567, 0);
+ vst1q_lane_s8(c1, vout0x01234567_1x01234567, 8);
+ }
+
+ nc = 0;
+ }
+ } while (nc != 0);
+}