Google Git
Sign in
android / platform / external / XNNPACK / refs/heads/main / . / src / amalgam / avx2.c
blob: a3e62ac373010be771edf3026caac9cc4746ada4 [file] [log] [blame] [edit]
// 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 <immintrin.h>
#include <xnnpack/common.h>
#include <xnnpack/dwconv.h>
#include <xnnpack/gemm.h>
#include <xnnpack/igemm.h>
#include <xnnpack/intrinsics-polyfill.h>
#include <xnnpack/lut.h>
#include <xnnpack/math.h>
#include <xnnpack/pavgpool.h>
#include <xnnpack/raddstoreexpminusmax.h>
#include <xnnpack/unaligned.h>
#include <xnnpack/vadd.h>
#include <xnnpack/vcvt.h>
#include <xnnpack/vlrelu.h>
#include <xnnpack/vunary.h>
void xnn_f16_gemm_minmax_ukernel_1x16__avx2_broadcast(
size_t mr,
size_t nc,
size_t kc,
const void*restrict a,
size_t a_stride,
const void*restrict w,
void*restrict c,
size_t cm_stride,
size_t cn_stride,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint16_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
const uint16_t* a0 = a;
uint16_t* c0 = c;
do {
__m256 vacc0x01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
__m256 vacc0x89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
size_t k = kc;
do {
const __m256 va0 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a0));
a0 += 1;
const __m256 vb01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
const __m256 vb89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
vacc0x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb01234567, vacc0x01234567), _MM_FROUND_NO_EXC));
vacc0x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb89ABCDEF, vacc0x89ABCDEF), _MM_FROUND_NO_EXC));
k -= sizeof(uint16_t);
} while (k != 0);
const __m256 vmin = _mm256_load_ps(params->avx.min);
vacc0x01234567 = _mm256_max_ps(vacc0x01234567, vmin);
vacc0x89ABCDEF = _mm256_max_ps(vacc0x89ABCDEF, vmin);
const __m256 vmax = _mm256_load_ps(params->avx.max);
vacc0x01234567 = _mm256_min_ps(vacc0x01234567, vmax);
vacc0x89ABCDEF = _mm256_min_ps(vacc0x89ABCDEF, vmax);
if XNN_LIKELY(nc >= 16) {
_mm_storeu_si128((__m128i*) c0, _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c0 + 8), _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC));
c0 = (uint16_t*) ((uintptr_t) c0 + cn_stride);
a0 = (const uint16_t*) ((uintptr_t) a0 - kc);
nc -= 16;
} else {
__m128i vh0x01234567 = _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC);
if (nc & 8) {
_mm_storeu_si128((__m128i*) c0, vh0x01234567);
vh0x01234567 = _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC);
c0 += 8;
}
if (nc & 4) {
_mm_storel_epi64((__m128i*) c0, vh0x01234567);
vh0x01234567 = _mm_unpackhi_epi64(vh0x01234567, vh0x01234567);
c0 += 4;
}
if (nc & 2) {
_mm_storeu_si32(c0, vh0x01234567);
vh0x01234567 = _mm_srli_epi64(vh0x01234567, 32);
c0 += 2;
}
if (nc & 1) {
*c0 = (uint16_t) _mm_extract_epi16(vh0x01234567, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_f16_gemm_minmax_ukernel_4x16__avx2_broadcast(
size_t mr,
size_t nc,
size_t kc,
const void*restrict a,
size_t a_stride,
const void*restrict w,
void*restrict c,
size_t cm_stride,
size_t cn_stride,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(mr != 0);
assert(mr <= 4);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint16_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
const uint16_t* a0 = a;
uint16_t* c0 = c;
const uint16_t* a1 = (const uint16_t*) ((uintptr_t) a0 + a_stride);
uint16_t* c1 = (uint16_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
a1 = a0;
c1 = c0;
}
const uint16_t* a2 = (const uint16_t*) ((uintptr_t) a1 + a_stride);
uint16_t* c2 = (uint16_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
a2 = a1;
c2 = c1;
}
const uint16_t* a3 = (const uint16_t*) ((uintptr_t) a2 + a_stride);
uint16_t* c3 = (uint16_t*) ((uintptr_t) c2 + cm_stride);
if XNN_UNPREDICTABLE(mr != 4) {
a3 = a2;
c3 = c2;
}
do {
__m256 vacc0x01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
__m256 vacc0x89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
__m256 vacc1x01234567 = vacc0x01234567;
__m256 vacc1x89ABCDEF = vacc0x89ABCDEF;
__m256 vacc2x01234567 = vacc0x01234567;
__m256 vacc2x89ABCDEF = vacc0x89ABCDEF;
__m256 vacc3x01234567 = vacc0x01234567;
__m256 vacc3x89ABCDEF = vacc0x89ABCDEF;
w = (const uint16_t*) w + 16;
size_t k = kc;
do {
const __m256 va0 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a0));
a0 += 1;
const __m256 va1 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a1));
a1 += 1;
const __m256 va2 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a2));
a2 += 1;
const __m256 va3 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a3));
a3 += 1;
const __m256 vb01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
const __m256 vb89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
vacc0x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb01234567, vacc0x01234567), _MM_FROUND_NO_EXC));
vacc1x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va1, vb01234567, vacc1x01234567), _MM_FROUND_NO_EXC));
vacc2x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va2, vb01234567, vacc2x01234567), _MM_FROUND_NO_EXC));
vacc3x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va3, vb01234567, vacc3x01234567), _MM_FROUND_NO_EXC));
vacc0x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb89ABCDEF, vacc0x89ABCDEF), _MM_FROUND_NO_EXC));
vacc1x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va1, vb89ABCDEF, vacc1x89ABCDEF), _MM_FROUND_NO_EXC));
vacc2x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va2, vb89ABCDEF, vacc2x89ABCDEF), _MM_FROUND_NO_EXC));
vacc3x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va3, vb89ABCDEF, vacc3x89ABCDEF), _MM_FROUND_NO_EXC));
k -= sizeof(uint16_t);
} while (k != 0);
const __m256 vmin = _mm256_load_ps(params->avx.min);
vacc0x01234567 = _mm256_max_ps(vacc0x01234567, vmin);
vacc1x01234567 = _mm256_max_ps(vacc1x01234567, vmin);
vacc2x01234567 = _mm256_max_ps(vacc2x01234567, vmin);
vacc3x01234567 = _mm256_max_ps(vacc3x01234567, vmin);
vacc0x89ABCDEF = _mm256_max_ps(vacc0x89ABCDEF, vmin);
vacc1x89ABCDEF = _mm256_max_ps(vacc1x89ABCDEF, vmin);
vacc2x89ABCDEF = _mm256_max_ps(vacc2x89ABCDEF, vmin);
vacc3x89ABCDEF = _mm256_max_ps(vacc3x89ABCDEF, vmin);
const __m256 vmax = _mm256_load_ps(params->avx.max);
vacc0x01234567 = _mm256_min_ps(vacc0x01234567, vmax);
vacc1x01234567 = _mm256_min_ps(vacc1x01234567, vmax);
vacc2x01234567 = _mm256_min_ps(vacc2x01234567, vmax);
vacc3x01234567 = _mm256_min_ps(vacc3x01234567, vmax);
vacc0x89ABCDEF = _mm256_min_ps(vacc0x89ABCDEF, vmax);
vacc1x89ABCDEF = _mm256_min_ps(vacc1x89ABCDEF, vmax);
vacc2x89ABCDEF = _mm256_min_ps(vacc2x89ABCDEF, vmax);
vacc3x89ABCDEF = _mm256_min_ps(vacc3x89ABCDEF, vmax);
if XNN_LIKELY(nc >= 16) {
_mm_storeu_si128((__m128i*) c0, _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c0 + 8), _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC));
c0 = (uint16_t*) ((uintptr_t) c0 + cn_stride);
_mm_storeu_si128((__m128i*) c1, _mm256_cvtps_ph(vacc1x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c1 + 8), _mm256_cvtps_ph(vacc1x89ABCDEF, _MM_FROUND_NO_EXC));
c1 = (uint16_t*) ((uintptr_t) c1 + cn_stride);
_mm_storeu_si128((__m128i*) c2, _mm256_cvtps_ph(vacc2x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c2 + 8), _mm256_cvtps_ph(vacc2x89ABCDEF, _MM_FROUND_NO_EXC));
c2 = (uint16_t*) ((uintptr_t) c2 + cn_stride);
_mm_storeu_si128((__m128i*) c3, _mm256_cvtps_ph(vacc3x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c3 + 8), _mm256_cvtps_ph(vacc3x89ABCDEF, _MM_FROUND_NO_EXC));
c3 = (uint16_t*) ((uintptr_t) c3 + cn_stride);
a0 = (const uint16_t*) ((uintptr_t) a0 - kc);
a1 = (const uint16_t*) ((uintptr_t) a1 - kc);
a2 = (const uint16_t*) ((uintptr_t) a2 - kc);
a3 = (const uint16_t*) ((uintptr_t) a3 - kc);
nc -= 16;
} else {
__m128i vh0x01234567 = _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC);
__m128i vh1x01234567 = _mm256_cvtps_ph(vacc1x01234567, _MM_FROUND_NO_EXC);
__m128i vh2x01234567 = _mm256_cvtps_ph(vacc2x01234567, _MM_FROUND_NO_EXC);
__m128i vh3x01234567 = _mm256_cvtps_ph(vacc3x01234567, _MM_FROUND_NO_EXC);
if (nc & 8) {
_mm_storeu_si128((__m128i*) c0, vh0x01234567);
_mm_storeu_si128((__m128i*) c1, vh1x01234567);
_mm_storeu_si128((__m128i*) c2, vh2x01234567);
_mm_storeu_si128((__m128i*) c3, vh3x01234567);
vh0x01234567 = _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC);
vh1x01234567 = _mm256_cvtps_ph(vacc1x89ABCDEF, _MM_FROUND_NO_EXC);
vh2x01234567 = _mm256_cvtps_ph(vacc2x89ABCDEF, _MM_FROUND_NO_EXC);
vh3x01234567 = _mm256_cvtps_ph(vacc3x89ABCDEF, _MM_FROUND_NO_EXC);
c0 += 8;
c1 += 8;
c2 += 8;
c3 += 8;
}
if (nc & 4) {
_mm_storel_epi64((__m128i*) c0, vh0x01234567);
_mm_storel_epi64((__m128i*) c1, vh1x01234567);
_mm_storel_epi64((__m128i*) c2, vh2x01234567);
_mm_storel_epi64((__m128i*) c3, vh3x01234567);
vh0x01234567 = _mm_unpackhi_epi64(vh0x01234567, vh0x01234567);
vh1x01234567 = _mm_unpackhi_epi64(vh1x01234567, vh1x01234567);
vh2x01234567 = _mm_unpackhi_epi64(vh2x01234567, vh2x01234567);
vh3x01234567 = _mm_unpackhi_epi64(vh3x01234567, vh3x01234567);
c0 += 4;
c1 += 4;
c2 += 4;
c3 += 4;
}
if (nc & 2) {
_mm_storeu_si32(c0, vh0x01234567);
_mm_storeu_si32(c1, vh1x01234567);
_mm_storeu_si32(c2, vh2x01234567);
_mm_storeu_si32(c3, vh3x01234567);
vh0x01234567 = _mm_srli_epi64(vh0x01234567, 32);
vh1x01234567 = _mm_srli_epi64(vh1x01234567, 32);
vh2x01234567 = _mm_srli_epi64(vh2x01234567, 32);
vh3x01234567 = _mm_srli_epi64(vh3x01234567, 32);
c0 += 2;
c1 += 2;
c2 += 2;
c3 += 2;
}
if (nc & 1) {
*c0 = (uint16_t) _mm_extract_epi16(vh0x01234567, 0);
*c1 = (uint16_t) _mm_extract_epi16(vh1x01234567, 0);
*c2 = (uint16_t) _mm_extract_epi16(vh2x01234567, 0);
*c3 = (uint16_t) _mm_extract_epi16(vh3x01234567, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_f16_igemm_minmax_ukernel_1x16__avx2_broadcast(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const void**restrict a,
const void*restrict w,
void*restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const void* zero,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint16_t) == 0);
assert(ks != 0);
assert(ks % (1 * sizeof(void*)) == 0);
assert(a_offset % sizeof(uint16_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
uint16_t* c0 = c;
do {
__m256 vacc0x01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
__m256 vacc0x89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
size_t p = ks;
do {
const uint16_t* restrict a0 = (const uint16_t*) a[0];
assert(a0 != NULL);
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const uint16_t*) ((uintptr_t) a0 + a_offset);
}
a += 1;
size_t k = kc;
do {
const __m256 vb01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
const __m256 vb89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
const __m256 va0 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a0));
a0 += 1;
vacc0x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb01234567, vacc0x01234567), _MM_FROUND_NO_EXC));
vacc0x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb89ABCDEF, vacc0x89ABCDEF), _MM_FROUND_NO_EXC));
k -= sizeof(uint16_t);
} while (k != 0);
p -= 1 * sizeof(void*);
} while (p != 0);
const __m256 vmin = _mm256_load_ps(params->avx.min);
vacc0x01234567 = _mm256_max_ps(vacc0x01234567, vmin);
vacc0x89ABCDEF = _mm256_max_ps(vacc0x89ABCDEF, vmin);
const __m256 vmax = _mm256_load_ps(params->avx.max);
vacc0x01234567 = _mm256_min_ps(vacc0x01234567, vmax);
vacc0x89ABCDEF = _mm256_min_ps(vacc0x89ABCDEF, vmax);
if XNN_LIKELY(nc >= 16) {
_mm_storeu_si128((__m128i*) c0, _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c0 + 8), _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC));
c0 = (uint16_t*) ((uintptr_t) c0 + cn_stride);
a = (const void**restrict) ((uintptr_t) a - ks);
nc -= 16;
} else {
__m128i vh0x01234567 = _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC);
if (nc & 8) {
_mm_storeu_si128((__m128i*) c0, vh0x01234567);
vh0x01234567 = _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC);
c0 += 8;
}
if (nc & 4) {
_mm_storel_epi64((__m128i*) c0, vh0x01234567);
vh0x01234567 = _mm_unpackhi_epi64(vh0x01234567, vh0x01234567);
c0 += 4;
}
if (nc & 2) {
_mm_storeu_si32(c0, vh0x01234567);
vh0x01234567 = _mm_srli_epi64(vh0x01234567, 32);
c0 += 2;
}
if (nc & 1) {
*c0 = _mm_extract_epi16(vh0x01234567, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_f16_igemm_minmax_ukernel_4x16__avx2_broadcast(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const void**restrict a,
const void*restrict w,
void*restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const void* zero,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(mr != 0);
assert(mr <= 4);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint16_t) == 0);
assert(ks != 0);
assert(ks % (4 * sizeof(void*)) == 0);
assert(a_offset % sizeof(uint16_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
uint16_t* c0 = c;
uint16_t* c1 = (uint16_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
c1 = c0;
}
uint16_t* c2 = (uint16_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
c2 = c1;
}
uint16_t* c3 = (uint16_t*) ((uintptr_t) c2 + cm_stride);
if XNN_UNPREDICTABLE(mr != 4) {
c3 = c2;
}
do {
__m256 vacc0x01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
__m256 vacc0x89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
__m256 vacc1x01234567 = vacc0x01234567;
__m256 vacc1x89ABCDEF = vacc0x89ABCDEF;
__m256 vacc2x01234567 = vacc0x01234567;
__m256 vacc2x89ABCDEF = vacc0x89ABCDEF;
__m256 vacc3x01234567 = vacc0x01234567;
__m256 vacc3x89ABCDEF = vacc0x89ABCDEF;
w = (const uint16_t*) w + 16;
size_t p = ks;
do {
const uint16_t* restrict a0 = (const uint16_t*) a[0];
assert(a0 != NULL);
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const uint16_t*) ((uintptr_t) a0 + a_offset);
}
const uint16_t* restrict a1 = (const uint16_t*) a[1];
assert(a1 != NULL);
if XNN_UNPREDICTABLE(a1 != zero) {
a1 = (const uint16_t*) ((uintptr_t) a1 + a_offset);
}
const uint16_t* restrict a2 = (const uint16_t*) a[2];
assert(a2 != NULL);
if XNN_UNPREDICTABLE(a2 != zero) {
a2 = (const uint16_t*) ((uintptr_t) a2 + a_offset);
}
const uint16_t* restrict a3 = (const uint16_t*) a[3];
assert(a3 != NULL);
if XNN_UNPREDICTABLE(a3 != zero) {
a3 = (const uint16_t*) ((uintptr_t) a3 + a_offset);
}
a += 4;
size_t k = kc;
do {
const __m256 vb01234567 = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) w));
const __m256 vb89ABCDEF = _mm256_cvtph_ps(_mm_load_si128((const __m128i*) ((const uint16_t*) w + 8)));
w = (const uint16_t*) w + 16;
const __m256 va0 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a0));
a0 += 1;
const __m256 va1 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a1));
a1 += 1;
const __m256 va2 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a2));
a2 += 1;
const __m256 va3 = _mm256_cvtph_ps(_mm_set1_epi16((short) *a3));
a3 += 1;
vacc0x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb01234567, vacc0x01234567), _MM_FROUND_NO_EXC));
vacc0x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va0, vb89ABCDEF, vacc0x89ABCDEF), _MM_FROUND_NO_EXC));
vacc1x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va1, vb01234567, vacc1x01234567), _MM_FROUND_NO_EXC));
vacc1x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va1, vb89ABCDEF, vacc1x89ABCDEF), _MM_FROUND_NO_EXC));
vacc2x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va2, vb01234567, vacc2x01234567), _MM_FROUND_NO_EXC));
vacc2x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va2, vb89ABCDEF, vacc2x89ABCDEF), _MM_FROUND_NO_EXC));
vacc3x01234567 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va3, vb01234567, vacc3x01234567), _MM_FROUND_NO_EXC));
vacc3x89ABCDEF = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_fmadd_ps(va3, vb89ABCDEF, vacc3x89ABCDEF), _MM_FROUND_NO_EXC));
k -= sizeof(uint16_t);
} while (k != 0);
p -= 4 * sizeof(void*);
} while (p != 0);
const __m256 vmin = _mm256_load_ps(params->avx.min);
vacc0x01234567 = _mm256_max_ps(vacc0x01234567, vmin);
vacc1x01234567 = _mm256_max_ps(vacc1x01234567, vmin);
vacc2x01234567 = _mm256_max_ps(vacc2x01234567, vmin);
vacc3x01234567 = _mm256_max_ps(vacc3x01234567, vmin);
vacc0x89ABCDEF = _mm256_max_ps(vacc0x89ABCDEF, vmin);
vacc1x89ABCDEF = _mm256_max_ps(vacc1x89ABCDEF, vmin);
vacc2x89ABCDEF = _mm256_max_ps(vacc2x89ABCDEF, vmin);
vacc3x89ABCDEF = _mm256_max_ps(vacc3x89ABCDEF, vmin);
const __m256 vmax = _mm256_load_ps(params->avx.max);
vacc0x01234567 = _mm256_min_ps(vacc0x01234567, vmax);
vacc1x01234567 = _mm256_min_ps(vacc1x01234567, vmax);
vacc2x01234567 = _mm256_min_ps(vacc2x01234567, vmax);
vacc3x01234567 = _mm256_min_ps(vacc3x01234567, vmax);
vacc0x89ABCDEF = _mm256_min_ps(vacc0x89ABCDEF, vmax);
vacc1x89ABCDEF = _mm256_min_ps(vacc1x89ABCDEF, vmax);
vacc2x89ABCDEF = _mm256_min_ps(vacc2x89ABCDEF, vmax);
vacc3x89ABCDEF = _mm256_min_ps(vacc3x89ABCDEF, vmax);
if XNN_LIKELY(nc >= 16) {
_mm_storeu_si128((__m128i*) c3, _mm256_cvtps_ph(vacc3x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c3 + 8), _mm256_cvtps_ph(vacc3x89ABCDEF, _MM_FROUND_NO_EXC));
c3 = (uint16_t*) ((uintptr_t) c3 + cn_stride);
_mm_storeu_si128((__m128i*) c2, _mm256_cvtps_ph(vacc2x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c2 + 8), _mm256_cvtps_ph(vacc2x89ABCDEF, _MM_FROUND_NO_EXC));
c2 = (uint16_t*) ((uintptr_t) c2 + cn_stride);
_mm_storeu_si128((__m128i*) c1, _mm256_cvtps_ph(vacc1x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c1 + 8), _mm256_cvtps_ph(vacc1x89ABCDEF, _MM_FROUND_NO_EXC));
c1 = (uint16_t*) ((uintptr_t) c1 + cn_stride);
_mm_storeu_si128((__m128i*) c0, _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (c0 + 8), _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC));
c0 = (uint16_t*) ((uintptr_t) c0 + cn_stride);
a = (const void**restrict) ((uintptr_t) a - ks);
nc -= 16;
} else {
__m128i vh3x01234567 = _mm256_cvtps_ph(vacc3x01234567, _MM_FROUND_NO_EXC);
__m128i vh2x01234567 = _mm256_cvtps_ph(vacc2x01234567, _MM_FROUND_NO_EXC);
__m128i vh1x01234567 = _mm256_cvtps_ph(vacc1x01234567, _MM_FROUND_NO_EXC);
__m128i vh0x01234567 = _mm256_cvtps_ph(vacc0x01234567, _MM_FROUND_NO_EXC);
if (nc & 8) {
_mm_storeu_si128((__m128i*) c3, vh3x01234567);
_mm_storeu_si128((__m128i*) c2, vh2x01234567);
_mm_storeu_si128((__m128i*) c1, vh1x01234567);
_mm_storeu_si128((__m128i*) c0, vh0x01234567);
vh3x01234567 = _mm256_cvtps_ph(vacc3x89ABCDEF, _MM_FROUND_NO_EXC);
vh2x01234567 = _mm256_cvtps_ph(vacc2x89ABCDEF, _MM_FROUND_NO_EXC);
vh1x01234567 = _mm256_cvtps_ph(vacc1x89ABCDEF, _MM_FROUND_NO_EXC);
vh0x01234567 = _mm256_cvtps_ph(vacc0x89ABCDEF, _MM_FROUND_NO_EXC);
c3 += 8;
c2 += 8;
c1 += 8;
c0 += 8;
}
if (nc & 4) {
_mm_storel_epi64((__m128i*) c3, vh3x01234567);
_mm_storel_epi64((__m128i*) c2, vh2x01234567);
_mm_storel_epi64((__m128i*) c1, vh1x01234567);
_mm_storel_epi64((__m128i*) c0, vh0x01234567);
vh3x01234567 = _mm_unpackhi_epi64(vh3x01234567, vh3x01234567);
vh2x01234567 = _mm_unpackhi_epi64(vh2x01234567, vh2x01234567);
vh1x01234567 = _mm_unpackhi_epi64(vh1x01234567, vh1x01234567);
vh0x01234567 = _mm_unpackhi_epi64(vh0x01234567, vh0x01234567);
c3 += 4;
c2 += 4;
c1 += 4;
c0 += 4;
}
if (nc & 2) {
_mm_storeu_si32(c3, vh3x01234567);
_mm_storeu_si32(c2, vh2x01234567);
_mm_storeu_si32(c1, vh1x01234567);
_mm_storeu_si32(c0, vh0x01234567);
vh3x01234567 = _mm_srli_epi64(vh3x01234567, 32);
vh2x01234567 = _mm_srli_epi64(vh2x01234567, 32);
vh1x01234567 = _mm_srli_epi64(vh1x01234567, 32);
vh0x01234567 = _mm_srli_epi64(vh0x01234567, 32);
c3 += 2;
c2 += 2;
c1 += 2;
c0 += 2;
}
if (nc & 1) {
*c3 = _mm_extract_epi16(vh3x01234567, 0);
*c2 = _mm_extract_epi16(vh2x01234567, 0);
*c1 = _mm_extract_epi16(vh1x01234567, 0);
*c0 = _mm_extract_epi16(vh0x01234567, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_f16_pavgpool_minmax_ukernel_9p8x__avx2_c8(
size_t output_pixels,
size_t kernel_elements,
size_t channels,
const void** input,
size_t input_offset,
const void* zero,
const void* multiplier,
void* buffer,
void* output,
size_t input_increment,
size_t output_increment,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(output_pixels != 0);
assert(kernel_elements > 9);
assert(channels != 0);
const __m256 voutput_min = _mm256_load_ps(params->avx.min);
const __m256 voutput_max = _mm256_load_ps(params->avx.max);
uint16_t* o = (uint16_t*) output;
do {
{
const uint16_t* i0 = (const uint16_t*) *input++;
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint16_t*) ((uintptr_t) i0 + input_offset);
}
const uint16_t* i1 = (const uint16_t*) *input++;
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint16_t*) ((uintptr_t) i1 + input_offset);
}
const uint16_t* i2 = (const uint16_t*) *input++;
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint16_t*) ((uintptr_t) i2 + input_offset);
}
const uint16_t* i3 = (const uint16_t*) *input++;
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint16_t*) ((uintptr_t) i3 + input_offset);
}
const uint16_t* i4 = (const uint16_t*) *input++;
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint16_t*) ((uintptr_t) i4 + input_offset);
}
const uint16_t* i5 = (const uint16_t*) *input++;
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint16_t*) ((uintptr_t) i5 + input_offset);
}
const uint16_t* i6 = (const uint16_t*) *input++;
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint16_t*) ((uintptr_t) i6 + input_offset);
}
const uint16_t* i7 = (const uint16_t*) *input++;
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint16_t*) ((uintptr_t) i7 + input_offset);
}
const uint16_t* i8 = (const uint16_t*) *input++;
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const uint16_t*) ((uintptr_t) i8 + input_offset);
}
uint16_t* b = (uint16_t*) buffer;
for (size_t c = 0; c < channels; c += 8) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
i0 += 8;
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
i1 += 8;
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
i2 += 8;
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
i3 += 8;
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
i4 += 8;
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
i5 += 8;
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
i6 += 8;
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
i7 += 8;
const __m256 vi8 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i8));
i8 += 8;
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum018 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vi8), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum01678 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum018, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum01678), _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) b, _mm256_cvtps_ph(vsum, _MM_FROUND_NO_EXC));
b += 8;
}
}
size_t k = kernel_elements;
for (k -= 9; k > 8; k -= 8) {
const uint16_t* i0 = (const uint16_t*) *input++;
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint16_t*) ((uintptr_t) i0 + input_offset);
}
const uint16_t* i1 = (const uint16_t*) *input++;
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint16_t*) ((uintptr_t) i1 + input_offset);
}
const uint16_t* i2 = (const uint16_t*) *input++;
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint16_t*) ((uintptr_t) i2 + input_offset);
}
const uint16_t* i3 = (const uint16_t*) *input++;
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint16_t*) ((uintptr_t) i3 + input_offset);
}
const uint16_t* i4 = (const uint16_t*) *input++;
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint16_t*) ((uintptr_t) i4 + input_offset);
}
const uint16_t* i5 = (const uint16_t*) *input++;
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint16_t*) ((uintptr_t) i5 + input_offset);
}
const uint16_t* i6 = (const uint16_t*) *input++;
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint16_t*) ((uintptr_t) i6 + input_offset);
}
const uint16_t* i7 = (const uint16_t*) *input++;
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint16_t*) ((uintptr_t) i7 + input_offset);
}
uint16_t* b = (uint16_t*) buffer;
for (size_t c = 0; c < channels; c += 8) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
i0 += 8;
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
i1 += 8;
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
i2 += 8;
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
i3 += 8;
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
i4 += 8;
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
i5 += 8;
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
i6 += 8;
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
i7 += 8;
const __m256 vacc = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) b));
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum01a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vacc), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum0167a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01a, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum0167a), _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) b, _mm256_cvtps_ph(vsum, _MM_FROUND_NO_EXC));
b += 8;
}
}
{
const uint16_t* i0 = (const uint16_t*) input[0];
assert(i0 != NULL);
const uint16_t* i1 = (const uint16_t*) input[1];
const uint16_t* i2 = (const uint16_t*) input[2];
const uint16_t* i3 = (const uint16_t*) input[3];
const uint16_t* i4 = (const uint16_t*) input[4];
const uint16_t* i5 = (const uint16_t*) input[5];
const uint16_t* i6 = (const uint16_t*) input[6];
const uint16_t* i7 = (const uint16_t*) input[7];
input = (const void**) ((uintptr_t) input + input_increment);
if (k < 2) {
i1 = (const uint16_t*) zero;
}
assert(i1 != NULL);
if (k <= 2) {
i2 = (const uint16_t*) zero;
}
assert(i2 != NULL);
if (k < 4) {
i3 = (const uint16_t*) zero;
}
assert(i3 != NULL);
if (k <= 4) {
i4 = (const uint16_t*) zero;
}
assert(i4 != NULL);
if (k < 6) {
i5 = (const uint16_t*) zero;
}
assert(i5 != NULL);
if (k <= 6) {
i6 = (const uint16_t*) zero;
}
assert(i6 != NULL);
if (k < 8) {
i7 = (const uint16_t*) zero;
}
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint16_t*) ((uintptr_t) i0 + input_offset);
}
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint16_t*) ((uintptr_t) i1 + input_offset);
}
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint16_t*) ((uintptr_t) i2 + input_offset);
}
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint16_t*) ((uintptr_t) i3 + input_offset);
}
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint16_t*) ((uintptr_t) i4 + input_offset);
}
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint16_t*) ((uintptr_t) i5 + input_offset);
}
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint16_t*) ((uintptr_t) i6 + input_offset);
}
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint16_t*) ((uintptr_t) i7 + input_offset);
}
const __m256 vmultiplier = _mm256_cvtph_ps(_mm_set1_epi16((short) *((const uint16_t*) multiplier)));
multiplier = (const uint16_t*) multiplier + 1;
size_t c = channels;
const uint16_t* b = (const uint16_t*) buffer;
while (c >= 8) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
i0 += 8;
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
i1 += 8;
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
i2 += 8;
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
i3 += 8;
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
i4 += 8;
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
i5 += 8;
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
i6 += 8;
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
i7 += 8;
const __m256 vacc = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) b));
b += 8;
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum01a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vacc), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum0167a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01a, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum0167a), _MM_FROUND_NO_EXC));
__m256 vout = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_mul_ps(vsum, vmultiplier), _MM_FROUND_NO_EXC));
vout = _mm256_max_ps(vout, voutput_min);
vout = _mm256_min_ps(vout, voutput_max);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vout, _MM_FROUND_NO_EXC));
o += 8;
c -= 8;
}
if (c != 0) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
const __m256 vacc = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) b));
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum01a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vacc), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum0167a = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01a, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum0167a), _MM_FROUND_NO_EXC));
__m256 vout = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_mul_ps(vsum, vmultiplier), _MM_FROUND_NO_EXC));
vout = _mm256_max_ps(vout, voutput_min);
vout = _mm256_min_ps(vout, voutput_max);
__m128i vh = _mm256_cvtps_ph(vout, _MM_FROUND_NO_EXC);
if (c & 4) {
_mm_storel_epi64((__m128i*) o, vh);
vh = _mm_unpackhi_epi64(vh, vh);
o += 4;
}
if (c & 2) {
_mm_storeu_si32(o, vh);
vh = _mm_srli_epi64(vh, 32);
o += 2;
}
if (c & 1) {
*o = (uint16_t) _mm_extract_epi16(vh, 0);
o += 1;
}
}
}
o = (uint16_t*) ((uintptr_t) o + output_increment);
} while (--output_pixels != 0);
}
void xnn_f16_pavgpool_minmax_ukernel_9x__avx2_c8(
size_t output_pixels,
size_t kernel_elements,
size_t channels,
const void** input,
size_t input_offset,
const void* zero,
const void* multiplier,
void* output,
size_t input_increment,
size_t output_increment,
const union xnn_f16_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(output_pixels != 0);
assert(kernel_elements != 0);
assert(kernel_elements <= 9);
assert(channels != 0);
const __m256 voutput_min = _mm256_load_ps(params->avx.min);
const __m256 voutput_max = _mm256_load_ps(params->avx.max);
uint16_t* o = (uint16_t*) output;
do {
const uint16_t* i0 = (const uint16_t*) input[0];
assert(i0 != NULL);
const uint16_t* i1 = (const uint16_t*) input[1];
const uint16_t* i2 = (const uint16_t*) input[2];
const uint16_t* i3 = (const uint16_t*) input[3];
const uint16_t* i4 = (const uint16_t*) input[4];
const uint16_t* i5 = (const uint16_t*) input[5];
const uint16_t* i6 = (const uint16_t*) input[6];
const uint16_t* i7 = (const uint16_t*) input[7];
const uint16_t* i8 = (const uint16_t*) input[8];
input = (const void**) ((uintptr_t) input + input_increment);
if (kernel_elements < 2) {
i1 = (const uint16_t*) zero;
}
assert(i1 != NULL);
if (kernel_elements <= 2) {
i2 = (const uint16_t*) zero;
}
assert(i2 != NULL);
if (kernel_elements < 4) {
i3 = (const uint16_t*) zero;
}
assert(i3 != NULL);
if (kernel_elements <= 4) {
i4 = (const uint16_t*) zero;
}
assert(i4 != NULL);
if (kernel_elements < 6) {
i5 = (const uint16_t*) zero;
}
assert(i5 != NULL);
if (kernel_elements <= 6) {
i6 = (const uint16_t*) zero;
}
assert(i6 != NULL);
if (kernel_elements < 8) {
i7 = (const uint16_t*) zero;
}
assert(i7 != NULL);
if (kernel_elements <= 8) {
i8 = (const uint16_t*) zero;
}
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint16_t*) ((uintptr_t) i0 + input_offset);
}
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint16_t*) ((uintptr_t) i1 + input_offset);
}
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint16_t*) ((uintptr_t) i2 + input_offset);
}
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint16_t*) ((uintptr_t) i3 + input_offset);
}
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint16_t*) ((uintptr_t) i4 + input_offset);
}
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint16_t*) ((uintptr_t) i5 + input_offset);
}
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint16_t*) ((uintptr_t) i6 + input_offset);
}
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint16_t*) ((uintptr_t) i7 + input_offset);
}
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const uint16_t*) ((uintptr_t) i8 + input_offset);
}
const __m256 vmultiplier = _mm256_cvtph_ps(_mm_set1_epi16((short) *((const uint16_t*) multiplier)));
multiplier = (const uint16_t*) multiplier + 1;
size_t c = channels;
while (c >= 8) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
i0 += 8;
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
i1 += 8;
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
i2 += 8;
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
i3 += 8;
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
i4 += 8;
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
i5 += 8;
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
i6 += 8;
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
i7 += 8;
const __m256 vi8 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i8));
i8 += 8;
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum018 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vi8), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum01678 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum018, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum01678), _MM_FROUND_NO_EXC));
__m256 vout = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_mul_ps(vsum, vmultiplier), _MM_FROUND_NO_EXC));
vout = _mm256_max_ps(vout, voutput_min);
vout = _mm256_min_ps(vout, voutput_max);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vout, _MM_FROUND_NO_EXC));
o += 8;
c -= 8;
}
if (c != 0) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i0));
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i1));
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i2));
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i3));
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i4));
const __m256 vi5 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i5));
const __m256 vi6 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i6));
const __m256 vi7 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i7));
const __m256 vi8 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i8));
const __m256 vsum01 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi0, vi1), _MM_FROUND_NO_EXC));
const __m256 vsum23 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi2, vi3), _MM_FROUND_NO_EXC));
const __m256 vsum45 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi4, vi5), _MM_FROUND_NO_EXC));
const __m256 vsum67 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vi6, vi7), _MM_FROUND_NO_EXC));
const __m256 vsum018 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum01, vi8), _MM_FROUND_NO_EXC));
const __m256 vsum2345 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum23, vsum45), _MM_FROUND_NO_EXC));
const __m256 vsum01678 = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum018, vsum67), _MM_FROUND_NO_EXC));
const __m256 vsum = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_add_ps(vsum2345, vsum01678), _MM_FROUND_NO_EXC));
__m256 vout = _mm256_cvtph_ps(_mm256_cvtps_ph(_mm256_mul_ps(vsum, vmultiplier), _MM_FROUND_NO_EXC));
vout = _mm256_max_ps(vout, voutput_min);
vout = _mm256_min_ps(vout, voutput_max);
__m128i vh = _mm256_cvtps_ph(vout, _MM_FROUND_NO_EXC);
if (c & 4) {
_mm_storel_epi64((__m128i*) o, vh);
vh = _mm_unpackhi_epi64(vh, vh);
o += 4;
}
if (c & 2) {
_mm_storeu_si32(o, vh);
vh = _mm_srli_epi64(vh, 32);
o += 2;
}
if (c & 1) {
*o = (uint16_t) _mm_extract_epi16(vh, 0);
o += 1;
}
}
o = (uint16_t*) ((uintptr_t) o + output_increment);
} while (--output_pixels != 0);
}
void xnn_f16_raddstoreexpminusmax_ukernel__avx2_rr1_p2_x40(
size_t batch,
const void* input,
const void* max,
void* output,
void* sum,
const union xnn_f16_expminus_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(batch % sizeof(uint16_t) == 0);
const __m256 vi_max = _mm256_cvtph_ps(_mm_set1_epi16((short) *((const uint16_t*) max)));
const __m256 vlog2e = _mm256_load_ps(params->avx2_rr1_p2.log2e);
const __m256 vmagic_bias = _mm256_load_ps(params->avx2_rr1_p2.magic_bias);
const __m256 vminus_ln2 = _mm256_load_ps(params->avx2_rr1_p2.minus_ln2);
const __m256 vc2 = _mm256_load_ps(params->avx2_rr1_p2.c2);
const __m256 vc1 = _mm256_load_ps(params->avx2_rr1_p2.c1);
const __m256 vdenorm_cutoff = _mm256_load_ps(params->avx2_rr1_p2.denorm_cutoff);
const uint16_t* i = (const uint16_t*) input;
uint16_t* o = (uint16_t*) output;
__m256 vacc0 = _mm256_setzero_ps();
for (; batch >= 40 * sizeof(uint16_t); batch -= 40 * sizeof(uint16_t)) {
const __m256 vi0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
const __m256 vi1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 8)));
const __m256 vi2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 16)));
const __m256 vi3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 24)));
const __m256 vi4 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 32)));
i += 40;
const __m256 vx0 = _mm256_sub_ps(vi0, vi_max);
const __m256 vx1 = _mm256_sub_ps(vi1, vi_max);
const __m256 vx2 = _mm256_sub_ps(vi2, vi_max);
const __m256 vx3 = _mm256_sub_ps(vi3, vi_max);
const __m256 vx4 = _mm256_sub_ps(vi4, vi_max);
__m256 vn0 = _mm256_fmadd_ps(vx0, vlog2e, vmagic_bias);
__m256 vn1 = _mm256_fmadd_ps(vx1, vlog2e, vmagic_bias);
__m256 vn2 = _mm256_fmadd_ps(vx2, vlog2e, vmagic_bias);
__m256 vn3 = _mm256_fmadd_ps(vx3, vlog2e, vmagic_bias);
__m256 vn4 = _mm256_fmadd_ps(vx4, vlog2e, vmagic_bias);
const __m256 vs0 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn0), 23));
const __m256 vs1 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn1), 23));
const __m256 vs2 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn2), 23));
const __m256 vs3 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn3), 23));
const __m256 vs4 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn4), 23));
vn0 = _mm256_sub_ps(vn0, vmagic_bias);
vn1 = _mm256_sub_ps(vn1, vmagic_bias);
vn2 = _mm256_sub_ps(vn2, vmagic_bias);
vn3 = _mm256_sub_ps(vn3, vmagic_bias);
vn4 = _mm256_sub_ps(vn4, vmagic_bias);
__m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vx0);
__m256 vt1 = _mm256_fmadd_ps(vn1, vminus_ln2, vx1);
__m256 vt2 = _mm256_fmadd_ps(vn2, vminus_ln2, vx2);
__m256 vt3 = _mm256_fmadd_ps(vn3, vminus_ln2, vx3);
__m256 vt4 = _mm256_fmadd_ps(vn4, vminus_ln2, vx4);
const __m256 vp0 = _mm256_fmadd_ps(vc2, vt0, vc1);
const __m256 vp1 = _mm256_fmadd_ps(vc2, vt1, vc1);
const __m256 vp2 = _mm256_fmadd_ps(vc2, vt2, vc1);
const __m256 vp3 = _mm256_fmadd_ps(vc2, vt3, vc1);
const __m256 vp4 = _mm256_fmadd_ps(vc2, vt4, vc1);
vt0 = _mm256_mul_ps(vt0, vs0);
vt1 = _mm256_mul_ps(vt1, vs1);
vt2 = _mm256_mul_ps(vt2, vs2);
vt3 = _mm256_mul_ps(vt3, vs3);
vt4 = _mm256_mul_ps(vt4, vs4);
__m256 vf0 = _mm256_fmadd_ps(vt0, vp0, vs0);
__m256 vf1 = _mm256_fmadd_ps(vt1, vp1, vs1);
__m256 vf2 = _mm256_fmadd_ps(vt2, vp2, vs2);
__m256 vf3 = _mm256_fmadd_ps(vt3, vp3, vs3);
__m256 vf4 = _mm256_fmadd_ps(vt4, vp4, vs4);
vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vx0, vdenorm_cutoff, _CMP_LT_OS), vf0);
vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vx1, vdenorm_cutoff, _CMP_LT_OS), vf1);
vf2 = _mm256_andnot_ps(_mm256_cmp_ps(vx2, vdenorm_cutoff, _CMP_LT_OS), vf2);
vf3 = _mm256_andnot_ps(_mm256_cmp_ps(vx3, vdenorm_cutoff, _CMP_LT_OS), vf3);
vf4 = _mm256_andnot_ps(_mm256_cmp_ps(vx4, vdenorm_cutoff, _CMP_LT_OS), vf4);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vf0, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 16), _mm256_cvtps_ph(vf2, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 24), _mm256_cvtps_ph(vf3, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 32), _mm256_cvtps_ph(vf4, _MM_FROUND_NO_EXC));
o += 40;
vacc0 = _mm256_add_ps(vacc0, vf0);
vacc0 = _mm256_add_ps(vacc0, vf1);
vacc0 = _mm256_add_ps(vacc0, vf2);
vacc0 = _mm256_add_ps(vacc0, vf3);
vacc0 = _mm256_add_ps(vacc0, vf4);
}
__m256 vacc = vacc0;
for (; batch >= 8 * sizeof(uint16_t); batch -= 8 * sizeof(uint16_t)) {
const __m256 vi = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
i += 8;
const __m256 vx = _mm256_sub_ps(vi, vi_max);
__m256 vn = _mm256_fmadd_ps(vx, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vx);
const __m256 vp = _mm256_fmadd_ps(vc2, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
__m256 vf = _mm256_fmadd_ps(vt, vp, vs);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vx, vdenorm_cutoff, _CMP_LT_OS), vf);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vf, _MM_FROUND_NO_EXC));
o += 8;
vacc = _mm256_add_ps(vacc, vf);
}
__m128 vacc_lo = _mm_add_ps(_mm256_castps256_ps128(vacc), _mm256_extractf128_ps(vacc, 1));
if (batch != 0) {
assert(batch >= 1 * sizeof(uint16_t));
assert(batch <= 7 * sizeof(uint16_t));
const __m256 vi = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
const __m256 vx = _mm256_sub_ps(vi, vi_max);
__m256 vn = _mm256_fmadd_ps(vx, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vx);
const __m256 vp = _mm256_fmadd_ps(vc2, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
__m256 vf = _mm256_fmadd_ps(vt, vp, vs);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vx, vdenorm_cutoff, _CMP_LT_OS), vf);
__m128i vh = _mm256_cvtps_ph(vf, _MM_FROUND_NO_EXC);
__m128 vf_lo = _mm256_castps256_ps128(vf);
if (batch & (4 * sizeof(uint16_t))) {
_mm_storel_epi64((__m128i*) o, vh);
vh = _mm_unpackhi_epi64(vh, vh);
vacc_lo = _mm_add_ps(vacc_lo, vf_lo);
vf_lo = _mm256_extractf128_ps(vf, 1);
o += 4;
}
if (batch & (2 * sizeof(uint16_t))) {
_mm_storeu_si32(o, vh);
vh = _mm_srli_epi64(vh, 32);
vacc_lo = _mm_blend_ps(_mm_add_ps(vacc_lo, vf_lo), vacc_lo, 0xC);
vf_lo = _mm_movehl_ps(vf_lo, vf_lo);
o += 2;
}
if (batch & (1 * sizeof(uint16_t))) {
*o = (uint16_t) _mm_extract_epi16(vh, 0);
vacc_lo = _mm_add_ss(vacc_lo, vf_lo);
}
}
vacc_lo = _mm_add_ps(vacc_lo, _mm_movehl_ps(vacc_lo, vacc_lo));
vacc_lo = _mm_add_ss(vacc_lo, _mm_movehdup_ps(vacc_lo));
*((uint16_t*) sum) = (uint16_t) _mm_extract_epi16(_mm_cvtps_ph(vacc_lo, _MM_FROUND_NO_EXC), 0);
_mm256_zeroupper();
}
void xnn_f16_velu_ukernel__avx2_rr1_p3_x16(
size_t n,
const void* input,
void* output,
const union xnn_f16_elu_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(n % sizeof(uint16_t) == 0);
const __m256 vprescale = _mm256_load_ps(params->avx2_rr1_p3.prescale);
const __m256 vsat_cutoff = _mm256_load_ps(params->avx2_rr1_p3.sat_cutoff);
const __m256 vmagic_bias = _mm256_load_ps(params->avx2_rr1_p3.magic_bias);
const __m256 vlog2e = _mm256_load_ps(params->avx2_rr1_p3.log2e);
const __m256 vminus_ln2 = _mm256_load_ps(params->avx2_rr1_p3.minus_ln2);
const __m256 vc3 = _mm256_load_ps(params->avx2_rr1_p3.c3);
const __m256 vc2 = _mm256_load_ps(params->avx2_rr1_p3.c2);
const __m256 vc1 = _mm256_load_ps(params->avx2_rr1_p3.c1);
const __m256 valpha = _mm256_load_ps(params->avx2_rr1_p3.alpha);
const __m256 vbeta = _mm256_load_ps(params->avx2_rr1_p3.beta);
const uint16_t* i = (const uint16_t*) input;
uint16_t* o = (uint16_t*) output;
for (; n >= 16 * sizeof(uint16_t); n -= 16 * sizeof(uint16_t)) {
__m256 vx0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
__m256 vx1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 8)));
i += 16;
const __m256 vz0 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx0, vprescale));
const __m256 vz1 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx1, vprescale));
__m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias);
__m256 vn1 = _mm256_fmadd_ps(vz1, vlog2e, vmagic_bias);
__m256 vs0 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn0), 23));
vn0 = _mm256_sub_ps(vn0, vmagic_bias);
__m256 vs1 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn1), 23));
vn1 = _mm256_sub_ps(vn1, vmagic_bias);
__m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0);
__m256 vt1 = _mm256_fmadd_ps(vn1, vminus_ln2, vz1);
__m256 vp0 = _mm256_fmadd_ps(vc3, vt0, vc2);
__m256 vp1 = _mm256_fmadd_ps(vc3, vt1, vc2);
vp0 = _mm256_fmadd_ps(vp0, vt0, vc1);
vt0 = _mm256_mul_ps(vt0, valpha);
vp1 = _mm256_fmadd_ps(vp1, vt1, vc1);
vt1 = _mm256_mul_ps(vt1, valpha);
vt0 = _mm256_mul_ps(vt0, vs0);
vs0 = _mm256_fmsub_ps(vs0, valpha, valpha);
vt1 = _mm256_mul_ps(vt1, vs1);
vs1 = _mm256_fmsub_ps(vs1, valpha, valpha);
const __m256 ve0 = _mm256_fmadd_ps(vp0, vt0, vs0);
vx0 = _mm256_mul_ps(vx0, vbeta);
const __m256 ve1 = _mm256_fmadd_ps(vp1, vt1, vs1);
vx1 = _mm256_mul_ps(vx1, vbeta);
const __m256 vy0 = _mm256_blendv_ps(vx0, ve0, vx0);
const __m256 vy1 = _mm256_blendv_ps(vx1, ve1, vx1);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vy0, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vy1, _MM_FROUND_NO_EXC));
o += 16;
}
for (; n >= 8 * sizeof(uint16_t); n -= 8 * sizeof(uint16_t)) {
__m256 vx = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
i += 8;
const __m256 vz = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx, vprescale));
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
__m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc3, vt, vc2);
vp = _mm256_fmadd_ps(vp, vt, vc1);
vt = _mm256_mul_ps(vt, valpha);
vt = _mm256_mul_ps(vt, vs);
vs = _mm256_fmsub_ps(vs, valpha, valpha);
const __m256 ve = _mm256_fmadd_ps(vp, vt, vs);
vx = _mm256_mul_ps(vx, vbeta);
const __m256 vy = _mm256_blendv_ps(vx, ve, vx);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vy, _MM_FROUND_NO_EXC));
o += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(uint16_t));
assert(n <= 7 * sizeof(uint16_t));
__m256 vx = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
const __m256 vz = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx, vprescale));
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
__m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc3, vt, vc2);
vp = _mm256_fmadd_ps(vp, vt, vc1);
vt = _mm256_mul_ps(vt, valpha);
vt = _mm256_mul_ps(vt, vs);
vs = _mm256_fmsub_ps(vs, valpha, valpha);
const __m256 ve = _mm256_fmadd_ps(vp, vt, vs);
vx = _mm256_mul_ps(vx, vbeta);
const __m256 vy = _mm256_blendv_ps(vx, ve, vx);
__m128i vh = _mm256_cvtps_ph(vy, _MM_FROUND_NO_EXC);
if (n & (4 * sizeof(uint16_t))) {
_mm_storel_epi64((__m128i*) o, vh);
vh = _mm_unpackhi_epi64(vh, vh);
o += 4;
}
if (n & (2 * sizeof(uint16_t))) {
_mm_storeu_si32(o, vh);
vh = _mm_srli_epi64(vh, 32);
o += 2;
}
if (n & (1 * sizeof(uint16_t))) {
*o = (uint16_t) _mm_extract_epi16(vh, 0);
}
}
}
void xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x32(
size_t batch,
const void* input,
void* output,
const union xnn_f16_sigmoid_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(batch % sizeof(uint16_t) == 0);
const __m256 vsign_mask = _mm256_load_ps(params->avx2_rr1_p2.sign_mask);
const __m256 vmagic_bias = _mm256_load_ps(params->avx2_rr1_p2.magic_bias);
const __m256 vlog2e = _mm256_load_ps(params->avx2_rr1_p2.log2e);
const __m256 vminus_ln2 = _mm256_load_ps(params->avx2_rr1_p2.minus_ln2);
const __m256 vc2 = _mm256_load_ps(params->avx2_rr1_p2.c2);
const __m256 vc1 = _mm256_load_ps(params->avx2_rr1_p2.c1);
const __m256 vone = _mm256_load_ps(params->avx2_rr1_p2.one);
const __m256 vdenorm_cutoff = _mm256_load_ps(params->avx2_rr1_p2.denorm_cutoff);
const uint16_t* i = (const uint16_t*) input;
uint16_t* o = (uint16_t*) output;
for (; batch >= 32 * sizeof(uint16_t); batch -= 32 * sizeof(uint16_t)) {
const __m256 vx0 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
const __m256 vx1 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 8)));
const __m256 vx2 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 16)));
const __m256 vx3 = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) (i + 24)));
i += 32;
const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask);
const __m256 vz1 = _mm256_or_ps(vx1, vsign_mask);
const __m256 vz2 = _mm256_or_ps(vx2, vsign_mask);
const __m256 vz3 = _mm256_or_ps(vx3, vsign_mask);
__m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias);
__m256 vn1 = _mm256_fmadd_ps(vz1, vlog2e, vmagic_bias);
__m256 vn2 = _mm256_fmadd_ps(vz2, vlog2e, vmagic_bias);
__m256 vn3 = _mm256_fmadd_ps(vz3, vlog2e, vmagic_bias);
const __m256 vs0 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn0), 23));
const __m256 vs1 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn1), 23));
const __m256 vs2 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn2), 23));
const __m256 vs3 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn3), 23));
vn0 = _mm256_sub_ps(vn0, vmagic_bias);
vn1 = _mm256_sub_ps(vn1, vmagic_bias);
vn2 = _mm256_sub_ps(vn2, vmagic_bias);
vn3 = _mm256_sub_ps(vn3, vmagic_bias);
__m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0);
__m256 vt1 = _mm256_fmadd_ps(vn1, vminus_ln2, vz1);
__m256 vt2 = _mm256_fmadd_ps(vn2, vminus_ln2, vz2);
__m256 vt3 = _mm256_fmadd_ps(vn3, vminus_ln2, vz3);
const __m256 vp0 = _mm256_fmadd_ps(vc2, vt0, vc1);
const __m256 vp1 = _mm256_fmadd_ps(vc2, vt1, vc1);
const __m256 vp2 = _mm256_fmadd_ps(vc2, vt2, vc1);
const __m256 vp3 = _mm256_fmadd_ps(vc2, vt3, vc1);
vt0 = _mm256_mul_ps(vt0, vs0);
vt1 = _mm256_mul_ps(vt1, vs1);
vt2 = _mm256_mul_ps(vt2, vs2);
vt3 = _mm256_mul_ps(vt3, vs3);
const __m256 ve0 = _mm256_fmadd_ps(vt0, vp0, vs0);
const __m256 ve1 = _mm256_fmadd_ps(vt1, vp1, vs1);
const __m256 ve2 = _mm256_fmadd_ps(vt2, vp2, vs2);
const __m256 ve3 = _mm256_fmadd_ps(vt3, vp3, vs3);
const __m256 vd0 = _mm256_add_ps(ve0, vone);
const __m256 vd1 = _mm256_add_ps(ve1, vone);
const __m256 vd2 = _mm256_add_ps(ve2, vone);
const __m256 vd3 = _mm256_add_ps(ve3, vone);
const __m256 vr0 = _mm256_rcp_ps(vd0);
const __m256 vr1 = _mm256_rcp_ps(vd1);
const __m256 vr2 = _mm256_rcp_ps(vd2);
const __m256 vr3 = _mm256_rcp_ps(vd3);
__m256 vf0 = _mm256_mul_ps(ve0, vr0);
__m256 vf1 = _mm256_mul_ps(ve1, vr1);
__m256 vf2 = _mm256_mul_ps(ve2, vr2);
__m256 vf3 = _mm256_mul_ps(ve3, vr3);
vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0);
vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1);
vf2 = _mm256_andnot_ps(_mm256_cmp_ps(vz2, vdenorm_cutoff, _CMP_LT_OS), vf2);
vf3 = _mm256_andnot_ps(_mm256_cmp_ps(vz3, vdenorm_cutoff, _CMP_LT_OS), vf3);
vf0 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf0), vf0, vx0);
vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1);
vf2 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf2), vf2, vx2);
vf3 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf3), vf3, vx3);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vf0, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 16), _mm256_cvtps_ph(vf2, _MM_FROUND_NO_EXC));
_mm_storeu_si128((__m128i*) (o + 24), _mm256_cvtps_ph(vf3, _MM_FROUND_NO_EXC));
o += 32;
}
for (; batch >= 8 * sizeof(uint16_t); batch -= 8 * sizeof(uint16_t)) {
const __m256 vx = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
i += 8;
const __m256 vz = _mm256_or_ps(vx, vsign_mask);
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
const __m256 vp = _mm256_fmadd_ps(vc2, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
const __m256 ve = _mm256_fmadd_ps(vt, vp, vs);
const __m256 vd = _mm256_add_ps(ve, vone);
const __m256 vr = _mm256_rcp_ps(vd);
__m256 vf = _mm256_mul_ps(ve, vr);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vz, vdenorm_cutoff, _CMP_LT_OS), vf);
vf = _mm256_blendv_ps(_mm256_sub_ps(vone, vf), vf, vx);
_mm_storeu_si128((__m128i*) o, _mm256_cvtps_ph(vf, _MM_FROUND_NO_EXC));
o += 8;
}
if XNN_UNLIKELY(batch != 0) {
assert(batch >= 1 * sizeof(uint16_t));
assert(batch <= 7 * sizeof(uint16_t));
const __m256 vx = _mm256_cvtph_ps(_mm_loadu_si128((const __m128i*) i));
const __m256 vz = _mm256_or_ps(vx, vsign_mask);
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
const __m256 vp = _mm256_fmadd_ps(vc2, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
const __m256 ve = _mm256_fmadd_ps(vt, vp, vs);
const __m256 vd = _mm256_add_ps(ve, vone);
const __m256 vr = _mm256_rcp_ps(vd);
__m256 vf = _mm256_mul_ps(ve, vr);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vz, vdenorm_cutoff, _CMP_LT_OS), vf);
vf = _mm256_blendv_ps(_mm256_sub_ps(vone, vf), vf, vx);
__m128i vh = _mm256_cvtps_ph(vf, _MM_FROUND_NO_EXC);
if (batch & (4 * sizeof(uint16_t))) {
_mm_storel_epi64((__m128i*) o, vh);
vh = _mm_unpackhi_epi64(vh, vh);
o += 4;
}
if (batch & (2 * sizeof(uint16_t))) {
_mm_storeu_si32(o, vh);
vh = _mm_srli_epi64(vh, 32);
o += 2;
}
if (batch & (1 * sizeof(uint16_t))) {
*o = (uint16_t) _mm_extract_epi16(vh, 0);
}
}
}
void xnn_f32_qs8_vcvt_ukernel__avx2_x64(
size_t n,
const float* x,
int8_t* y,
const union xnn_f32_qs8_cvt_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256 vscale = _mm256_load_ps(params->avx2.scale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->avx2.output_max_less_zero_point);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m256i vshuffle_mask = _mm256_load_si256((const __m256i*) params->avx2.shuffle_mask);
const __m256i voutput_min = _mm256_load_si256((const __m256i*) params->avx2.output_min);
for (; n >= 64 * sizeof(float); n -= 64 * sizeof(float)) {
__m256 vx01 = _mm256_loadu_ps(x);
__m256 vx23 = _mm256_loadu_ps(x + 8);
__m256 vx45 = _mm256_loadu_ps(x + 16);
__m256 vx67 = _mm256_loadu_ps(x + 24);
__m256 vx89 = _mm256_loadu_ps(x + 32);
__m256 vxAB = _mm256_loadu_ps(x + 40);
__m256 vxCD = _mm256_loadu_ps(x + 48);
__m256 vxEF = _mm256_loadu_ps(x + 56);
x += 64;
vx01 = _mm256_mul_ps(vx01, vscale);
vx23 = _mm256_mul_ps(vx23, vscale);
vx45 = _mm256_mul_ps(vx45, vscale);
vx67 = _mm256_mul_ps(vx67, vscale);
vx89 = _mm256_mul_ps(vx89, vscale);
vxAB = _mm256_mul_ps(vxAB, vscale);
vxCD = _mm256_mul_ps(vxCD, vscale);
vxEF = _mm256_mul_ps(vxEF, vscale);
vx01 = _mm256_min_ps(vx01, voutput_max_less_zero_point);
vx23 = _mm256_min_ps(vx23, voutput_max_less_zero_point);
vx45 = _mm256_min_ps(vx45, voutput_max_less_zero_point);
vx67 = _mm256_min_ps(vx67, voutput_max_less_zero_point);
vx89 = _mm256_min_ps(vx89, voutput_max_less_zero_point);
vxAB = _mm256_min_ps(vxAB, voutput_max_less_zero_point);
vxCD = _mm256_min_ps(vxCD, voutput_max_less_zero_point);
vxEF = _mm256_min_ps(vxEF, voutput_max_less_zero_point);
const __m256i vacc01 = _mm256_cvtps_epi32(vx01);
const __m256i vacc23 = _mm256_cvtps_epi32(vx23);
const __m256i vacc45 = _mm256_cvtps_epi32(vx45);
const __m256i vacc67 = _mm256_cvtps_epi32(vx67);
const __m256i vacc89 = _mm256_cvtps_epi32(vx89);
const __m256i vaccAB = _mm256_cvtps_epi32(vxAB);
const __m256i vaccCD = _mm256_cvtps_epi32(vxCD);
const __m256i vaccEF = _mm256_cvtps_epi32(vxEF);
__m256i vacc0213 = _mm256_packs_epi32(vacc01, vacc23);
__m256i vacc4657 = _mm256_packs_epi32(vacc45, vacc67);
__m256i vacc8A9B = _mm256_packs_epi32(vacc89, vaccAB);
__m256i vaccCEDF = _mm256_packs_epi32(vaccCD, vaccEF);
vacc0213 = _mm256_adds_epi16(vacc0213, voutput_zero_point);
vacc4657 = _mm256_adds_epi16(vacc4657, voutput_zero_point);
vacc8A9B = _mm256_adds_epi16(vacc8A9B, voutput_zero_point);
vaccCEDF = _mm256_adds_epi16(vaccCEDF, voutput_zero_point);
const __m256i vy02461357 = _mm256_packs_epi16(vacc0213, vacc4657);
const __m256i vy8ACE9BDF = _mm256_packs_epi16(vacc8A9B, vaccCEDF);
__m256i vy01234567 = _mm256_permutevar8x32_epi32(vy02461357, vshuffle_mask);
__m256i vy89ABCDEF = _mm256_permutevar8x32_epi32(vy8ACE9BDF, vshuffle_mask);
vy01234567 = _mm256_max_epi8(vy01234567, voutput_min);
vy89ABCDEF = _mm256_max_epi8(vy89ABCDEF, voutput_min);
_mm256_storeu_si256((__m256i*) y, vy01234567);
_mm256_storeu_si256((__m256i*) (y + 32), vy89ABCDEF);
y += 64;
}
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
__m256 vx = _mm256_loadu_ps(x);
vx = _mm256_mul_ps(vx, vscale);
vx = _mm256_min_ps(vx, voutput_max_less_zero_point);
x += 8;
const __m256i vacc = _mm256_cvtps_epi32(vx);
__m128i vy = _mm_packs_epi32(_mm256_castsi256_si128(vacc), _mm256_extracti128_si256(vacc, 1));
vy = _mm_adds_epi16(vy, _mm256_castsi256_si128(voutput_zero_point));
vy = _mm_packs_epi16(vy, vy);
vy = _mm_max_epi8(vy, _mm256_castsi256_si128(voutput_min));
_mm_storel_epi64((__m128i*) y, vy);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(float));
assert(n <= 7 * sizeof(float));
const __m256i vmask = _mm256_loadu_si256((const __m256i*) ((uintptr_t) &params->avx2.mask_table[7] - n));
__m256 vx = _mm256_maskload_ps(x, vmask);
vx = _mm256_mul_ps(vx, vscale);
vx = _mm256_min_ps(vx, voutput_max_less_zero_point);
const __m256i vacc = _mm256_cvtps_epi32(vx);
__m128i vy = _mm_packs_epi32(_mm256_castsi256_si128(vacc), _mm256_extracti128_si256(vacc, 1));
vy = _mm_adds_epi16(vy, _mm256_castsi256_si128(voutput_zero_point));
vy = _mm_packs_epi16(vy, vy);
vy = _mm_max_epi8(vy, _mm256_castsi256_si128(voutput_min));
if (n & (4 * sizeof(float))) {
_mm_storeu_si32(y, vy);
y += 4;
vy = _mm_srli_epi64(vy, 32);
}
if (n & (2 * sizeof(float))) {
_mm_storeu_si16(y, vy);
y += 2;
vy = _mm_srli_epi32(vy, 16);
}
if (n & (1 * sizeof(float))) {
*y = (int8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_f32_qu8_vcvt_ukernel__avx2_x64(
size_t n,
const float* x,
uint8_t* y,
const union xnn_f32_qu8_cvt_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(n != 0);
assert(n % sizeof(float) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256 vscale = _mm256_load_ps(params->avx2.scale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->avx2.output_max_less_zero_point);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m256i vshuffle_mask = _mm256_load_si256((const __m256i*) params->avx2.shuffle_mask);
const __m256i voutput_min = _mm256_load_si256((const __m256i*) params->avx2.output_min);
for (; n >= 64 * sizeof(float); n -= 64 * sizeof(float)) {
__m256 vx01 = _mm256_loadu_ps(x);
__m256 vx23 = _mm256_loadu_ps(x + 8);
__m256 vx45 = _mm256_loadu_ps(x + 16);
__m256 vx67 = _mm256_loadu_ps(x + 24);
__m256 vx89 = _mm256_loadu_ps(x + 32);
__m256 vxAB = _mm256_loadu_ps(x + 40);
__m256 vxCD = _mm256_loadu_ps(x + 48);
__m256 vxEF = _mm256_loadu_ps(x + 56);
x += 64;
vx01 = _mm256_mul_ps(vx01, vscale);
vx23 = _mm256_mul_ps(vx23, vscale);
vx45 = _mm256_mul_ps(vx45, vscale);
vx67 = _mm256_mul_ps(vx67, vscale);
vx89 = _mm256_mul_ps(vx89, vscale);
vxAB = _mm256_mul_ps(vxAB, vscale);
vxCD = _mm256_mul_ps(vxCD, vscale);
vxEF = _mm256_mul_ps(vxEF, vscale);
vx01 = _mm256_min_ps(vx01, voutput_max_less_zero_point);
vx23 = _mm256_min_ps(vx23, voutput_max_less_zero_point);
vx45 = _mm256_min_ps(vx45, voutput_max_less_zero_point);
vx67 = _mm256_min_ps(vx67, voutput_max_less_zero_point);
vx89 = _mm256_min_ps(vx89, voutput_max_less_zero_point);
vxAB = _mm256_min_ps(vxAB, voutput_max_less_zero_point);
vxCD = _mm256_min_ps(vxCD, voutput_max_less_zero_point);
vxEF = _mm256_min_ps(vxEF, voutput_max_less_zero_point);
const __m256i vacc01 = _mm256_cvtps_epi32(vx01);
const __m256i vacc23 = _mm256_cvtps_epi32(vx23);
const __m256i vacc45 = _mm256_cvtps_epi32(vx45);
const __m256i vacc67 = _mm256_cvtps_epi32(vx67);
const __m256i vacc89 = _mm256_cvtps_epi32(vx89);
const __m256i vaccAB = _mm256_cvtps_epi32(vxAB);
const __m256i vaccCD = _mm256_cvtps_epi32(vxCD);
const __m256i vaccEF = _mm256_cvtps_epi32(vxEF);
__m256i vacc0213 = _mm256_packs_epi32(vacc01, vacc23);
__m256i vacc4657 = _mm256_packs_epi32(vacc45, vacc67);
__m256i vacc8A9B = _mm256_packs_epi32(vacc89, vaccAB);
__m256i vaccCEDF = _mm256_packs_epi32(vaccCD, vaccEF);
vacc0213 = _mm256_adds_epi16(vacc0213, voutput_zero_point);
vacc4657 = _mm256_adds_epi16(vacc4657, voutput_zero_point);
vacc8A9B = _mm256_adds_epi16(vacc8A9B, voutput_zero_point);
vaccCEDF = _mm256_adds_epi16(vaccCEDF, voutput_zero_point);
const __m256i vy02461357 = _mm256_packus_epi16(vacc0213, vacc4657);
const __m256i vy8ACE9BDF = _mm256_packus_epi16(vacc8A9B, vaccCEDF);
__m256i vy01234567 = _mm256_permutevar8x32_epi32(vy02461357, vshuffle_mask);
__m256i vy89ABCDEF = _mm256_permutevar8x32_epi32(vy8ACE9BDF, vshuffle_mask);
vy01234567 = _mm256_max_epu8(vy01234567, voutput_min);
vy89ABCDEF = _mm256_max_epu8(vy89ABCDEF, voutput_min);
_mm256_storeu_si256((__m256i*) y, vy01234567);
_mm256_storeu_si256((__m256i*) (y + 32), vy89ABCDEF);
y += 64;
}
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
__m256 vx = _mm256_loadu_ps(x);
vx = _mm256_mul_ps(vx, vscale);
vx = _mm256_min_ps(vx, voutput_max_less_zero_point);
x += 8;
const __m256i vacc = _mm256_cvtps_epi32(vx);
__m128i vy = _mm_packs_epi32(_mm256_castsi256_si128(vacc), _mm256_extracti128_si256(vacc, 1));
vy = _mm_adds_epi16(vy, _mm256_castsi256_si128(voutput_zero_point));
vy = _mm_packus_epi16(vy, vy);
vy = _mm_max_epu8(vy, _mm256_castsi256_si128(voutput_min));
_mm_storel_epi64((__m128i*) y, vy);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(float));
assert(n <= 7 * sizeof(float));
const __m256i vmask = _mm256_loadu_si256((const __m256i*) ((uintptr_t) &params->avx2.mask_table[7] - n));
__m256 vx = _mm256_maskload_ps(x, vmask);
vx = _mm256_mul_ps(vx, vscale);
vx = _mm256_min_ps(vx, voutput_max_less_zero_point);
const __m256i vacc = _mm256_cvtps_epi32(vx);
__m128i vy = _mm_packs_epi32(_mm256_castsi256_si128(vacc), _mm256_extracti128_si256(vacc, 1));
vy = _mm_adds_epi16(vy, _mm256_castsi256_si128(voutput_zero_point));
vy = _mm_packus_epi16(vy, vy);
vy = _mm_max_epu8(vy, _mm256_castsi256_si128(voutput_min));
if (n & (4 * sizeof(float))) {
_mm_storeu_si32(y, vy);
y += 4;
vy = _mm_srli_epi64(vy, 32);
}
if (n & (2 * sizeof(float))) {
_mm_storeu_si16(y, vy);
y += 2;
vy = _mm_srli_epi32(vy, 16);
}
if (n & (1 * sizeof(float))) {
*y = (uint8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_f32_velu_ukernel__avx2_rr1_lut4_p4_perm_x56(
size_t n,
const float* x,
float* y,
const union xnn_f32_elu_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(n % sizeof(float) == 0);
const __m256 vprescale = _mm256_load_ps(params->avx2_rr1_lut4_p4.prescale);
const __m256 valpha = _mm256_load_ps(params->avx2_rr1_lut4_p4.alpha);
const __m256 vbeta = _mm256_load_ps(params->avx2_rr1_lut4_p4.beta);
const __m256 vsat_cutoff = _mm256_load_ps(params->avx2_rr1_lut4_p4.sat_cutoff);
const __m256 vmagic_bias = _mm256_load_ps(params->avx2_rr1_lut4_p4.magic_bias);
const __m256 vlog2e = _mm256_load_ps(params->avx2_rr1_lut4_p4.log2e);
const __m256 vtable = _mm256_load_ps(params->avx2_rr1_lut4_p4.table);
const __m256 vminus_ln2 = _mm256_load_ps(params->avx2_rr1_lut4_p4.minus_ln2);
const __m256 vc4 = _mm256_load_ps(params->avx2_rr1_lut4_p4.c4);
const __m256 vc3 = _mm256_load_ps(params->avx2_rr1_lut4_p4.c3);
const __m256 vc2 = _mm256_load_ps(params->avx2_rr1_lut4_p4.c2);
for (; n >= 56 * sizeof(float); n -= 56 * sizeof(float)) {
__m256 vx0 = _mm256_loadu_ps(x);
__m256 vx1 = _mm256_loadu_ps(x + 8);
__m256 vx2 = _mm256_loadu_ps(x + 16);
__m256 vx3 = _mm256_loadu_ps(x + 24);
__m256 vx4 = _mm256_loadu_ps(x + 32);
__m256 vx5 = _mm256_loadu_ps(x + 40);
__m256 vx6 = _mm256_loadu_ps(x + 48);
x += 56;
const __m256 vz0 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx0, vprescale));
const __m256 vz1 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx1, vprescale));
const __m256 vz2 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx2, vprescale));
const __m256 vz3 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx3, vprescale));
const __m256 vz4 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx4, vprescale));
const __m256 vz5 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx5, vprescale));
const __m256 vz6 = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx6, vprescale));
__m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias);
__m256 vn1 = _mm256_fmadd_ps(vz1, vlog2e, vmagic_bias);
__m256 vn2 = _mm256_fmadd_ps(vz2, vlog2e, vmagic_bias);
__m256 vn3 = _mm256_fmadd_ps(vz3, vlog2e, vmagic_bias);
__m256 vn4 = _mm256_fmadd_ps(vz4, vlog2e, vmagic_bias);
__m256 vn5 = _mm256_fmadd_ps(vz5, vlog2e, vmagic_bias);
__m256 vn6 = _mm256_fmadd_ps(vz6, vlog2e, vmagic_bias);
const __m256i ven0 = _mm256_slli_epi32(_mm256_castps_si256(vn0), 21);
const __m256i vl0 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn0)));
vn0 = _mm256_sub_ps(vn0, vmagic_bias);
const __m256i ven1 = _mm256_slli_epi32(_mm256_castps_si256(vn1), 21);
const __m256i vl1 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn1)));
vn1 = _mm256_sub_ps(vn1, vmagic_bias);
const __m256i ven2 = _mm256_slli_epi32(_mm256_castps_si256(vn2), 21);
const __m256i vl2 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn2)));
vn2 = _mm256_sub_ps(vn2, vmagic_bias);
const __m256i ven3 = _mm256_slli_epi32(_mm256_castps_si256(vn3), 21);
const __m256i vl3 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn3)));
vn3 = _mm256_sub_ps(vn3, vmagic_bias);
const __m256i ven4 = _mm256_slli_epi32(_mm256_castps_si256(vn4), 21);
const __m256i vl4 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn4)));
vn4 = _mm256_sub_ps(vn4, vmagic_bias);
const __m256i ven5 = _mm256_slli_epi32(_mm256_castps_si256(vn5), 21);
const __m256i vl5 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn5)));
vn5 = _mm256_sub_ps(vn5, vmagic_bias);
const __m256i ven6 = _mm256_slli_epi32(_mm256_castps_si256(vn6), 21);
const __m256i vl6 = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn6)));
vn6 = _mm256_sub_ps(vn6, vmagic_bias);
__m256 vs0 = _mm256_castsi256_ps(_mm256_add_epi32(vl0, ven0));
__m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0);
__m256 vs1 = _mm256_castsi256_ps(_mm256_add_epi32(vl1, ven1));
__m256 vt1 = _mm256_fmadd_ps(vn1, vminus_ln2, vz1);
__m256 vs2 = _mm256_castsi256_ps(_mm256_add_epi32(vl2, ven2));
__m256 vt2 = _mm256_fmadd_ps(vn2, vminus_ln2, vz2);
__m256 vs3 = _mm256_castsi256_ps(_mm256_add_epi32(vl3, ven3));
__m256 vt3 = _mm256_fmadd_ps(vn3, vminus_ln2, vz3);
__m256 vs4 = _mm256_castsi256_ps(_mm256_add_epi32(vl4, ven4));
__m256 vt4 = _mm256_fmadd_ps(vn4, vminus_ln2, vz4);
__m256 vs5 = _mm256_castsi256_ps(_mm256_add_epi32(vl5, ven5));
__m256 vt5 = _mm256_fmadd_ps(vn5, vminus_ln2, vz5);
__m256 vs6 = _mm256_castsi256_ps(_mm256_add_epi32(vl6, ven6));
__m256 vt6 = _mm256_fmadd_ps(vn6, vminus_ln2, vz6);
__m256 vp0 = _mm256_fmadd_ps(vc4, vt0, vc3);
__m256 vp1 = _mm256_fmadd_ps(vc4, vt1, vc3);
__m256 vp2 = _mm256_fmadd_ps(vc4, vt2, vc3);
__m256 vp3 = _mm256_fmadd_ps(vc4, vt3, vc3);
__m256 vp4 = _mm256_fmadd_ps(vc4, vt4, vc3);
__m256 vp5 = _mm256_fmadd_ps(vc4, vt5, vc3);
__m256 vp6 = _mm256_fmadd_ps(vc4, vt6, vc3);
vp0 = _mm256_fmadd_ps(vp0, vt0, vc2);
vp1 = _mm256_fmadd_ps(vp1, vt1, vc2);
vp2 = _mm256_fmadd_ps(vp2, vt2, vc2);
vp3 = _mm256_fmadd_ps(vp3, vt3, vc2);
vp4 = _mm256_fmadd_ps(vp4, vt4, vc2);
vp5 = _mm256_fmadd_ps(vp5, vt5, vc2);
vp6 = _mm256_fmadd_ps(vp6, vt6, vc2);
vp0 = _mm256_mul_ps(vp0, vt0);
vt0 = _mm256_mul_ps(vt0, vs0);
vp1 = _mm256_mul_ps(vp1, vt1);
vt1 = _mm256_mul_ps(vt1, vs1);
vp2 = _mm256_mul_ps(vp2, vt2);
vt2 = _mm256_mul_ps(vt2, vs2);
vp3 = _mm256_mul_ps(vp3, vt3);
vt3 = _mm256_mul_ps(vt3, vs3);
vp4 = _mm256_mul_ps(vp4, vt4);
vt4 = _mm256_mul_ps(vt4, vs4);
vp5 = _mm256_mul_ps(vp5, vt5);
vt5 = _mm256_mul_ps(vt5, vs5);
vp6 = _mm256_mul_ps(vp6, vt6);
vt6 = _mm256_mul_ps(vt6, vs6);
vs0 = _mm256_fmsub_ps(vs0, valpha, valpha);
vp0 = _mm256_fmadd_ps(vp0, vt0, vt0);
vs1 = _mm256_fmsub_ps(vs1, valpha, valpha);
vp1 = _mm256_fmadd_ps(vp1, vt1, vt1);
vs2 = _mm256_fmsub_ps(vs2, valpha, valpha);
vp2 = _mm256_fmadd_ps(vp2, vt2, vt2);
vs3 = _mm256_fmsub_ps(vs3, valpha, valpha);
vp3 = _mm256_fmadd_ps(vp3, vt3, vt3);
vs4 = _mm256_fmsub_ps(vs4, valpha, valpha);
vp4 = _mm256_fmadd_ps(vp4, vt4, vt4);
vs5 = _mm256_fmsub_ps(vs5, valpha, valpha);
vp5 = _mm256_fmadd_ps(vp5, vt5, vt5);
vs6 = _mm256_fmsub_ps(vs6, valpha, valpha);
vp6 = _mm256_fmadd_ps(vp6, vt6, vt6);
const __m256 ve0 = _mm256_fmadd_ps(vp0, valpha, vs0);
vx0 = _mm256_mul_ps(vx0, vbeta);
const __m256 ve1 = _mm256_fmadd_ps(vp1, valpha, vs1);
vx1 = _mm256_mul_ps(vx1, vbeta);
const __m256 ve2 = _mm256_fmadd_ps(vp2, valpha, vs2);
vx2 = _mm256_mul_ps(vx2, vbeta);
const __m256 ve3 = _mm256_fmadd_ps(vp3, valpha, vs3);
vx3 = _mm256_mul_ps(vx3, vbeta);
const __m256 ve4 = _mm256_fmadd_ps(vp4, valpha, vs4);
vx4 = _mm256_mul_ps(vx4, vbeta);
const __m256 ve5 = _mm256_fmadd_ps(vp5, valpha, vs5);
vx5 = _mm256_mul_ps(vx5, vbeta);
const __m256 ve6 = _mm256_fmadd_ps(vp6, valpha, vs6);
vx6 = _mm256_mul_ps(vx6, vbeta);
const __m256 vy0 = _mm256_blendv_ps(vx0, ve0, vx0);
const __m256 vy1 = _mm256_blendv_ps(vx1, ve1, vx1);
const __m256 vy2 = _mm256_blendv_ps(vx2, ve2, vx2);
const __m256 vy3 = _mm256_blendv_ps(vx3, ve3, vx3);
const __m256 vy4 = _mm256_blendv_ps(vx4, ve4, vx4);
const __m256 vy5 = _mm256_blendv_ps(vx5, ve5, vx5);
const __m256 vy6 = _mm256_blendv_ps(vx6, ve6, vx6);
_mm256_storeu_ps(y, vy0);
_mm256_storeu_ps(y + 8, vy1);
_mm256_storeu_ps(y + 16, vy2);
_mm256_storeu_ps(y + 24, vy3);
_mm256_storeu_ps(y + 32, vy4);
_mm256_storeu_ps(y + 40, vy5);
_mm256_storeu_ps(y + 48, vy6);
y += 56;
}
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
__m256 vx = _mm256_loadu_ps(x);
x += 8;
const __m256 vz = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx, vprescale));
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256i ven = _mm256_slli_epi32(_mm256_castps_si256(vn), 21);
const __m256i vl = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn)));
__m256 vs = _mm256_castsi256_ps(_mm256_add_epi32(vl, ven));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc4, vt, vc3);
vp = _mm256_fmadd_ps(vp, vt, vc2);
vp = _mm256_mul_ps(vp, vt);
vt = _mm256_mul_ps(vt, vs);
vs = _mm256_fmsub_ps(vs, valpha, valpha);
vp = _mm256_fmadd_ps(vp, vt, vt);
const __m256 ve = _mm256_fmadd_ps(vp, valpha, vs);
vx = _mm256_mul_ps(vx, vbeta);
const __m256 vy = _mm256_blendv_ps(vx, ve, vx);
_mm256_storeu_ps(y, vy);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(float));
assert(n <= 7 * sizeof(float));
const __m256i vmask = _mm256_loadu_si256((const __m256i*) ((uintptr_t) &params->avx2_rr1_lut4_p4.mask_table[7] - n));
__m256 vx = _mm256_maskload_ps(x, vmask);
const __m256 vz = _mm256_max_ps(vsat_cutoff, _mm256_mul_ps(vx, vprescale));
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256i ven = _mm256_slli_epi32(_mm256_castps_si256(vn), 21);
const __m256i vl = _mm256_castps_si256(_mm256_permutevar_ps(vtable, _mm256_castps_si256(vn)));
__m256 vs = _mm256_castsi256_ps(_mm256_add_epi32(vl, ven));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc4, vt, vc3);
vp = _mm256_fmadd_ps(vp, vt, vc2);
vp = _mm256_mul_ps(vp, vt);
vt = _mm256_mul_ps(vt, vs);
vs = _mm256_fmsub_ps(vs, valpha, valpha);
vp = _mm256_fmadd_ps(vp, vt, vt);
const __m256 ve = _mm256_fmadd_ps(vp, valpha, vs);
vx = _mm256_mul_ps(vx, vbeta);
const __m256 vy = _mm256_blendv_ps(vx, ve, vx);
__m128 vy_lo = _mm256_castps256_ps128(vy);
if (n & (4 * sizeof(float))) {
_mm_storeu_ps(y, vy_lo);
vy_lo = _mm256_extractf128_ps(vy, 1);
y += 4;
}
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vy_lo);
vy_lo = _mm_movehl_ps(vy_lo, vy_lo);
y += 2;
}
if (n & (1 * sizeof(float))) {
_mm_store_ss(y, vy_lo);
}
}
}
void xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x40(
size_t n,
const float* x,
float* y,
const union xnn_f32_sigmoid_params params[restrict XNN_MIN_ELEMENTS(1)])
{
assert(n % sizeof(float) == 0);
const __m256 vsign_mask = _mm256_load_ps(params->avx2_rr1_p5.sign_mask);
const __m256 vmagic_bias = _mm256_load_ps(params->avx2_rr1_p5.magic_bias);
const __m256 vlog2e = _mm256_load_ps(params->avx2_rr1_p5.log2e);
const __m256 vminus_ln2 = _mm256_load_ps(params->avx2_rr1_p5.minus_ln2);
const __m256 vc5 = _mm256_load_ps(params->avx2_rr1_p5.c5);
const __m256 vc4 = _mm256_load_ps(params->avx2_rr1_p5.c4);
const __m256 vc3 = _mm256_load_ps(params->avx2_rr1_p5.c3);
const __m256 vc2 = _mm256_load_ps(params->avx2_rr1_p5.c2);
const __m256 vc1 = _mm256_load_ps(params->avx2_rr1_p5.c1);
const __m256 vone = _mm256_load_ps(params->avx2_rr1_p5.one);
const __m256 vdenorm_cutoff = _mm256_load_ps(params->avx2_rr1_p5.denorm_cutoff);
for (; n >= 40 * sizeof(float); n -= 40 * sizeof(float)) {
const __m256 vx0 = _mm256_loadu_ps(x);
const __m256 vx1 = _mm256_loadu_ps(x + 8);
const __m256 vx2 = _mm256_loadu_ps(x + 16);
const __m256 vx3 = _mm256_loadu_ps(x + 24);
const __m256 vx4 = _mm256_loadu_ps(x + 32);
x += 40;
const __m256 vz0 = _mm256_or_ps(vx0, vsign_mask);
const __m256 vz1 = _mm256_or_ps(vx1, vsign_mask);
const __m256 vz2 = _mm256_or_ps(vx2, vsign_mask);
const __m256 vz3 = _mm256_or_ps(vx3, vsign_mask);
const __m256 vz4 = _mm256_or_ps(vx4, vsign_mask);
__m256 vn0 = _mm256_fmadd_ps(vz0, vlog2e, vmagic_bias);
__m256 vn1 = _mm256_fmadd_ps(vz1, vlog2e, vmagic_bias);
__m256 vn2 = _mm256_fmadd_ps(vz2, vlog2e, vmagic_bias);
__m256 vn3 = _mm256_fmadd_ps(vz3, vlog2e, vmagic_bias);
__m256 vn4 = _mm256_fmadd_ps(vz4, vlog2e, vmagic_bias);
const __m256 vs0 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn0), 23));
const __m256 vs1 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn1), 23));
const __m256 vs2 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn2), 23));
const __m256 vs3 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn3), 23));
const __m256 vs4 = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn4), 23));
vn0 = _mm256_sub_ps(vn0, vmagic_bias);
vn1 = _mm256_sub_ps(vn1, vmagic_bias);
vn2 = _mm256_sub_ps(vn2, vmagic_bias);
vn3 = _mm256_sub_ps(vn3, vmagic_bias);
vn4 = _mm256_sub_ps(vn4, vmagic_bias);
__m256 vt0 = _mm256_fmadd_ps(vn0, vminus_ln2, vz0);
__m256 vt1 = _mm256_fmadd_ps(vn1, vminus_ln2, vz1);
__m256 vt2 = _mm256_fmadd_ps(vn2, vminus_ln2, vz2);
__m256 vt3 = _mm256_fmadd_ps(vn3, vminus_ln2, vz3);
__m256 vt4 = _mm256_fmadd_ps(vn4, vminus_ln2, vz4);
__m256 vp0 = _mm256_fmadd_ps(vc5, vt0, vc4);
__m256 vp1 = _mm256_fmadd_ps(vc5, vt1, vc4);
__m256 vp2 = _mm256_fmadd_ps(vc5, vt2, vc4);
__m256 vp3 = _mm256_fmadd_ps(vc5, vt3, vc4);
__m256 vp4 = _mm256_fmadd_ps(vc5, vt4, vc4);
vp0 = _mm256_fmadd_ps(vp0, vt0, vc3);
vp1 = _mm256_fmadd_ps(vp1, vt1, vc3);
vp2 = _mm256_fmadd_ps(vp2, vt2, vc3);
vp3 = _mm256_fmadd_ps(vp3, vt3, vc3);
vp4 = _mm256_fmadd_ps(vp4, vt4, vc3);
vp0 = _mm256_fmadd_ps(vp0, vt0, vc2);
vp1 = _mm256_fmadd_ps(vp1, vt1, vc2);
vp2 = _mm256_fmadd_ps(vp2, vt2, vc2);
vp3 = _mm256_fmadd_ps(vp3, vt3, vc2);
vp4 = _mm256_fmadd_ps(vp4, vt4, vc2);
vp0 = _mm256_fmadd_ps(vp0, vt0, vc1);
vp1 = _mm256_fmadd_ps(vp1, vt1, vc1);
vp2 = _mm256_fmadd_ps(vp2, vt2, vc1);
vp3 = _mm256_fmadd_ps(vp3, vt3, vc1);
vp4 = _mm256_fmadd_ps(vp4, vt4, vc1);
vt0 = _mm256_mul_ps(vt0, vs0);
vt1 = _mm256_mul_ps(vt1, vs1);
vt2 = _mm256_mul_ps(vt2, vs2);
vt3 = _mm256_mul_ps(vt3, vs3);
vt4 = _mm256_mul_ps(vt4, vs4);
const __m256 ve0 = _mm256_fmadd_ps(vt0, vp0, vs0);
const __m256 ve1 = _mm256_fmadd_ps(vt1, vp1, vs1);
const __m256 ve2 = _mm256_fmadd_ps(vt2, vp2, vs2);
const __m256 ve3 = _mm256_fmadd_ps(vt3, vp3, vs3);
const __m256 ve4 = _mm256_fmadd_ps(vt4, vp4, vs4);
const __m256 vd0 = _mm256_add_ps(ve0, vone);
const __m256 vd1 = _mm256_add_ps(ve1, vone);
const __m256 vd2 = _mm256_add_ps(ve2, vone);
const __m256 vd3 = _mm256_add_ps(ve3, vone);
const __m256 vd4 = _mm256_add_ps(ve4, vone);
__m256 vf0 = _mm256_div_ps(ve0, vd0);
__m256 vf1 = _mm256_div_ps(ve1, vd1);
__m256 vf2 = _mm256_div_ps(ve2, vd2);
__m256 vf3 = _mm256_div_ps(ve3, vd3);
__m256 vf4 = _mm256_div_ps(ve4, vd4);
vf0 = _mm256_andnot_ps(_mm256_cmp_ps(vz0, vdenorm_cutoff, _CMP_LT_OS), vf0);
vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1);
vf2 = _mm256_andnot_ps(_mm256_cmp_ps(vz2, vdenorm_cutoff, _CMP_LT_OS), vf2);
vf3 = _mm256_andnot_ps(_mm256_cmp_ps(vz3, vdenorm_cutoff, _CMP_LT_OS), vf3);
vf4 = _mm256_andnot_ps(_mm256_cmp_ps(vz4, vdenorm_cutoff, _CMP_LT_OS), vf4);
vf0 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf0), vf0, vx0);
vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1);
vf2 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf2), vf2, vx2);
vf3 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf3), vf3, vx3);
vf4 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf4), vf4, vx4);
_mm256_storeu_ps(y, vf0);
_mm256_storeu_ps(y + 8, vf1);
_mm256_storeu_ps(y + 16, vf2);
_mm256_storeu_ps(y + 24, vf3);
_mm256_storeu_ps(y + 32, vf4);
y += 40;
}
for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
const __m256 vx = _mm256_loadu_ps(x);
x += 8;
const __m256 vz = _mm256_or_ps(vx, vsign_mask);
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc5, vt, vc4);
vp = _mm256_fmadd_ps(vp, vt, vc3);
vp = _mm256_fmadd_ps(vp, vt, vc2);
vp = _mm256_fmadd_ps(vp, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
const __m256 ve = _mm256_fmadd_ps(vt, vp, vs);
const __m256 vd = _mm256_add_ps(ve, vone);
__m256 vf = _mm256_div_ps(ve, vd);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vz, vdenorm_cutoff, _CMP_LT_OS), vf);
vf = _mm256_blendv_ps(_mm256_sub_ps(vone, vf), vf, vx);
_mm256_storeu_ps(y, vf);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(float));
assert(n <= 7 * sizeof(float));
const __m256i vmask = _mm256_loadu_si256((const __m256i*) ((uintptr_t) &params->avx2_rr1_p5.mask_table[7] - n));
const __m256 vx = _mm256_maskload_ps(x, vmask);
const __m256 vz = _mm256_or_ps(vx, vsign_mask);
__m256 vn = _mm256_fmadd_ps(vz, vlog2e, vmagic_bias);
const __m256 vs = _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_castps_si256(vn), 23));
vn = _mm256_sub_ps(vn, vmagic_bias);
__m256 vt = _mm256_fmadd_ps(vn, vminus_ln2, vz);
__m256 vp = _mm256_fmadd_ps(vc5, vt, vc4);
vp = _mm256_fmadd_ps(vp, vt, vc3);
vp = _mm256_fmadd_ps(vp, vt, vc2);
vp = _mm256_fmadd_ps(vp, vt, vc1);
vt = _mm256_mul_ps(vt, vs);
const __m256 ve = _mm256_fmadd_ps(vt, vp, vs);
const __m256 vd = _mm256_add_ps(ve, vone);
__m256 vf = _mm256_div_ps(ve, vd);
vf = _mm256_andnot_ps(_mm256_cmp_ps(vz, vdenorm_cutoff, _CMP_LT_OS), vf);
vf = _mm256_blendv_ps(_mm256_sub_ps(vone, vf), vf, vx);
__m128 vf_lo = _mm256_castps256_ps128(vf);
if (n & (4 * sizeof(float))) {
_mm_storeu_ps(y, vf_lo);
vf_lo = _mm256_extractf128_ps(vf, 1);
y += 4;
}
if (n & (2 * sizeof(float))) {
_mm_storel_pi((__m64*) y, vf_lo);
vf_lo = _mm_movehl_ps(vf_lo, vf_lo);
y += 2;
}
if (n & (1 * sizeof(float))) {
_mm_store_ss(y, vf_lo);
}
}
}
void xnn_qc8_dwconv_minmax_fp32_ukernel_up16x25__avx2_mul32(
size_t channels,
size_t output_width,
const int8_t** input,
const void* weights,
int8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const int8_t* zero,
const union xnn_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
do {
const int8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const int8_t*) ((uintptr_t) i0 + input_offset);
}
const int8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const int8_t*) ((uintptr_t) i1 + input_offset);
}
const int8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const int8_t*) ((uintptr_t) i2 + input_offset);
}
const int8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const int8_t*) ((uintptr_t) i3 + input_offset);
}
const int8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const int8_t*) ((uintptr_t) i4 + input_offset);
}
const int8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const int8_t*) ((uintptr_t) i5 + input_offset);
}
const int8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const int8_t*) ((uintptr_t) i6 + input_offset);
}
const int8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const int8_t*) ((uintptr_t) i7 + input_offset);
}
const int8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const int8_t*) ((uintptr_t) i8 + input_offset);
}
const int8_t* i9 = input[9];
assert(i9 != NULL);
if XNN_UNPREDICTABLE(i9 != zero) {
i9 = (const int8_t*) ((uintptr_t) i9 + input_offset);
}
const int8_t* i10 = input[10];
assert(i10 != NULL);
if XNN_UNPREDICTABLE(i10 != zero) {
i10 = (const int8_t*) ((uintptr_t) i10 + input_offset);
}
const int8_t* i11 = input[11];
assert(i11 != NULL);
if XNN_UNPREDICTABLE(i11 != zero) {
i11 = (const int8_t*) ((uintptr_t) i11 + input_offset);
}
const int8_t* i12 = input[12];
assert(i12 != NULL);
if XNN_UNPREDICTABLE(i12 != zero) {
i12 = (const int8_t*) ((uintptr_t) i12 + input_offset);
}
const int8_t* i13 = input[13];
assert(i13 != NULL);
if XNN_UNPREDICTABLE(i13 != zero) {
i13 = (const int8_t*) ((uintptr_t) i13 + input_offset);
}
const int8_t* i14 = input[14];
assert(i14 != NULL);
if XNN_UNPREDICTABLE(i14 != zero) {
i14 = (const int8_t*) ((uintptr_t) i14 + input_offset);
}
const int8_t* i15 = input[15];
assert(i15 != NULL);
if XNN_UNPREDICTABLE(i15 != zero) {
i15 = (const int8_t*) ((uintptr_t) i15 + input_offset);
}
const int8_t* i16 = input[16];
assert(i16 != NULL);
if XNN_UNPREDICTABLE(i16 != zero) {
i16 = (const int8_t*) ((uintptr_t) i16 + input_offset);
}
const int8_t* i17 = input[17];
assert(i17 != NULL);
if XNN_UNPREDICTABLE(i17 != zero) {
i17 = (const int8_t*) ((uintptr_t) i17 + input_offset);
}
const int8_t* i18 = input[18];
assert(i18 != NULL);
if XNN_UNPREDICTABLE(i18 != zero) {
i18 = (const int8_t*) ((uintptr_t) i18 + input_offset);
}
const int8_t* i19 = input[19];
assert(i19 != NULL);
if XNN_UNPREDICTABLE(i19 != zero) {
i19 = (const int8_t*) ((uintptr_t) i19 + input_offset);
}
const int8_t* i20 = input[20];
assert(i20 != NULL);
if XNN_UNPREDICTABLE(i20 != zero) {
i20 = (const int8_t*) ((uintptr_t) i20 + input_offset);
}
const int8_t* i21 = input[21];
assert(i21 != NULL);
if XNN_UNPREDICTABLE(i21 != zero) {
i21 = (const int8_t*) ((uintptr_t) i21 + input_offset);
}
const int8_t* i22 = input[22];
assert(i22 != NULL);
if XNN_UNPREDICTABLE(i22 != zero) {
i22 = (const int8_t*) ((uintptr_t) i22 + input_offset);
}
const int8_t* i23 = input[23];
assert(i23 != NULL);
if XNN_UNPREDICTABLE(i23 != zero) {
i23 = (const int8_t*) ((uintptr_t) i23 + input_offset);
}
const int8_t* i24 = input[24];
assert(i24 != NULL);
if XNN_UNPREDICTABLE(i24 != zero) {
i24 = (const int8_t*) ((uintptr_t) i24 + input_offset);
}
input = (const int8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(int8_t))));
const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(int8_t))));
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(int8_t))));
const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(int8_t))));
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(int8_t))));
const __m256i vi2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(int8_t))));
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t))));
const __m256i vi3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(int8_t))));
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(int8_t))));
const __m256i vi4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(int8_t))));
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(int8_t))));
const __m256i vi5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(int8_t))));
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(int8_t))));
const __m256i vi6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(int8_t))));
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(int8_t))));
const __m256i vi7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(int8_t))));
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(int8_t))));
const __m256i vi8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(int8_t))));
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
const __m256i vi9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(int8_t))));
const __m256i vi9x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i9 + 8)));
const __m256i vk9x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 152 * sizeof(int8_t))));
i9 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi9x89ABCDEF, vk9x89ABCDEF));
const __m256i vi10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 160 * sizeof(int8_t))));
const __m256i vi10x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i10 + 8)));
const __m256i vk10x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 168 * sizeof(int8_t))));
i10 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi10x89ABCDEF, vk10x89ABCDEF));
const __m256i vi11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 176 * sizeof(int8_t))));
const __m256i vi11x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i11 + 8)));
const __m256i vk11x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 184 * sizeof(int8_t))));
i11 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi11x89ABCDEF, vk11x89ABCDEF));
const __m256i vi12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 192 * sizeof(int8_t))));
const __m256i vi12x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i12 + 8)));
const __m256i vk12x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 200 * sizeof(int8_t))));
i12 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi12x89ABCDEF, vk12x89ABCDEF));
const __m256i vi13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 208 * sizeof(int8_t))));
const __m256i vi13x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i13 + 8)));
const __m256i vk13x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 216 * sizeof(int8_t))));
i13 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi13x89ABCDEF, vk13x89ABCDEF));
const __m256i vi14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 224 * sizeof(int8_t))));
const __m256i vi14x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i14 + 8)));
const __m256i vk14x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 232 * sizeof(int8_t))));
i14 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi14x89ABCDEF, vk14x89ABCDEF));
const __m256i vi15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 240 * sizeof(int8_t))));
const __m256i vi15x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i15 + 8)));
const __m256i vk15x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 248 * sizeof(int8_t))));
i15 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi15x89ABCDEF, vk15x89ABCDEF));
const __m256i vi16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 256 * sizeof(int8_t))));
const __m256i vi16x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i16 + 8)));
const __m256i vk16x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 264 * sizeof(int8_t))));
i16 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi16x89ABCDEF, vk16x89ABCDEF));
const __m256i vi17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 272 * sizeof(int8_t))));
const __m256i vi17x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i17 + 8)));
const __m256i vk17x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 280 * sizeof(int8_t))));
i17 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi17x89ABCDEF, vk17x89ABCDEF));
const __m256i vi18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 288 * sizeof(int8_t))));
const __m256i vi18x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i18 + 8)));
const __m256i vk18x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 296 * sizeof(int8_t))));
i18 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi18x89ABCDEF, vk18x89ABCDEF));
const __m256i vi19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 304 * sizeof(int8_t))));
const __m256i vi19x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i19 + 8)));
const __m256i vk19x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 312 * sizeof(int8_t))));
i19 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi19x89ABCDEF, vk19x89ABCDEF));
const __m256i vi20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 320 * sizeof(int8_t))));
const __m256i vi20x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i20 + 8)));
const __m256i vk20x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 328 * sizeof(int8_t))));
i20 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi20x89ABCDEF, vk20x89ABCDEF));
const __m256i vi21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 336 * sizeof(int8_t))));
const __m256i vi21x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i21 + 8)));
const __m256i vk21x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 344 * sizeof(int8_t))));
i21 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi21x89ABCDEF, vk21x89ABCDEF));
const __m256i vi22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 352 * sizeof(int8_t))));
const __m256i vi22x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i22 + 8)));
const __m256i vk22x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 360 * sizeof(int8_t))));
i22 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi22x89ABCDEF, vk22x89ABCDEF));
const __m256i vi23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 368 * sizeof(int8_t))));
const __m256i vi23x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i23 + 8)));
const __m256i vk23x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 376 * sizeof(int8_t))));
i23 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi23x89ABCDEF, vk23x89ABCDEF));
const __m256i vi24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 384 * sizeof(int8_t))));
const __m256i vi24x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i24 + 8)));
const __m256i vk24x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 392 * sizeof(int8_t))));
i24 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi24x89ABCDEF, vk24x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 400 * sizeof(int8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) w);
const __m256 vscale89ABCDEF = _mm256_loadu_ps((const float*) w + 8);
w = (const void*) ((const float*) w + 16);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale89ABCDEF);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const int8_t* k = (const int8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16)));
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32)));
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48)));
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64)));
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80)));
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96)));
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112)));
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128)));
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
const __m256i vi9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 144)));
i9 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
const __m256i vi10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 160)));
i10 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
const __m256i vi11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 176)));
i11 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
const __m256i vi12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 192)));
i12 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
const __m256i vi13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 208)));
i13 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
const __m256i vi14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 224)));
i14 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
const __m256i vi15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 240)));
i15 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
const __m256i vi16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 256)));
i16 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
const __m256i vi17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 272)));
i17 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
const __m256i vi18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 288)));
i18 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
const __m256i vi19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 304)));
i19 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
const __m256i vi20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 320)));
i20 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
const __m256i vi21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 336)));
i21 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
const __m256i vi22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 352)));
i22 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
const __m256i vi23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 368)));
i23 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
const __m256i vi24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 384)));
i24 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) ((uintptr_t) w + 16 * sizeof(int32_t) + 400 * sizeof(int8_t)));
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (int8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qc8_dwconv_minmax_fp32_ukernel_up16x3__avx2_mul32(
size_t channels,
size_t output_width,
const int8_t** input,
const void* weights,
int8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const int8_t* zero,
const union xnn_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
do {
const int8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const int8_t*) ((uintptr_t) i0 + input_offset);
}
const int8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const int8_t*) ((uintptr_t) i1 + input_offset);
}
const int8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const int8_t*) ((uintptr_t) i2 + input_offset);
}
input = (const int8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(int8_t))));
const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(int8_t))));
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(int8_t))));
const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(int8_t))));
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(int8_t))));
const __m256i vi2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(int8_t))));
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) w);
const __m256 vscale89ABCDEF = _mm256_loadu_ps((const float*) w + 8);
w = (const void*) ((const float*) w + 16);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale89ABCDEF);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const int8_t* k = (const int8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16)));
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32)));
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t)));
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (int8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qc8_dwconv_minmax_fp32_ukernel_up16x9__avx2_mul32(
size_t channels,
size_t output_width,
const int8_t** input,
const void* weights,
int8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const int8_t* zero,
const union xnn_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
do {
const int8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const int8_t*) ((uintptr_t) i0 + input_offset);
}
const int8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const int8_t*) ((uintptr_t) i1 + input_offset);
}
const int8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const int8_t*) ((uintptr_t) i2 + input_offset);
}
const int8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const int8_t*) ((uintptr_t) i3 + input_offset);
}
const int8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const int8_t*) ((uintptr_t) i4 + input_offset);
}
const int8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const int8_t*) ((uintptr_t) i5 + input_offset);
}
const int8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const int8_t*) ((uintptr_t) i6 + input_offset);
}
const int8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const int8_t*) ((uintptr_t) i7 + input_offset);
}
const int8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const int8_t*) ((uintptr_t) i8 + input_offset);
}
input = (const int8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(int8_t))));
const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(int8_t))));
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(int8_t))));
const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(int8_t))));
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(int8_t))));
const __m256i vi2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(int8_t))));
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t))));
const __m256i vi3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(int8_t))));
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(int8_t))));
const __m256i vi4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(int8_t))));
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(int8_t))));
const __m256i vi5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(int8_t))));
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(int8_t))));
const __m256i vi6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(int8_t))));
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(int8_t))));
const __m256i vi7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(int8_t))));
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(int8_t))));
const __m256i vi8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(int8_t))));
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(int8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) w);
const __m256 vscale89ABCDEF = _mm256_loadu_ps((const float*) w + 8);
w = (const void*) ((const float*) w + 16);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale89ABCDEF);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const int8_t* k = (const int8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16)));
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32)));
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48)));
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64)));
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80)));
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96)));
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112)));
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128)));
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
const __m256 vscale01234567 = _mm256_loadu_ps((const float*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(int8_t)));
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale01234567);
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (int8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qc8_gemm_minmax_fp32_ukernel_1x8c8__avx2(
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_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
const int8_t* a0 = a;
int8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale01234567 = _mm256_load_ps(w);
w = (const void*) ((const float*) w + 8);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale01234567);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a0 = (const int8_t*) ((uintptr_t) a0 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qc8_gemm_minmax_fp32_ukernel_3x8c8__avx2(
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_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
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;
}
const int8_t* a2 = (const int8_t*) ((uintptr_t) a1 + a_stride);
int8_t* c2 = (int8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
a2 = a1;
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepi8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepi8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale01234567 = _mm256_load_ps(w);
w = (const void*) ((const float*) w + 8);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale01234567);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale01234567);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale01234567);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
c1 = (int8_t*) ((uintptr_t) c1 + cn_stride);
c2 = (int8_t*) ((uintptr_t) c2 + cn_stride);
a0 = (const int8_t*) ((uintptr_t) a0 - kc);
a1 = (const int8_t*) ((uintptr_t) a1 - kc);
a2 = (const int8_t*) ((uintptr_t) a2 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
_mm_storeu_si32(c1, vout_hi);
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
c0 += 4;
c1 += 4;
c2 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
c0 += 2;
c1 += 2;
c2 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
*c1 = (int8_t) _mm_extract_epi8(vout_hi, 0);
*c2 = (int8_t) _mm_extract_epi8(vout_lo, 8);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qc8_igemm_minmax_fp32_ukernel_1x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const int8_t** restrict a,
const void* restrict w,
int8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const int8_t* zero,
const union xnn_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (1 * sizeof(void*)) == 0);
assert(a_offset % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
int8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t p = ks;
do {
const int8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const int8_t*) ((uintptr_t) a0 + a_offset);
}
a += 1;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
p -= 1 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale01234567 = _mm256_load_ps(w);
w = (const void*) ((const float*) w + 8);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale01234567);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a = (const int8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qc8_igemm_minmax_fp32_ukernel_3x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const int8_t** restrict a,
const void* restrict w,
int8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const int8_t* zero,
const union xnn_qc8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (3 * sizeof(void*)) == 0);
assert(a_offset % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
int8_t* c0 = c;
int8_t* c1 = (int8_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
c1 = c0;
}
int8_t* c2 = (int8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t p = ks;
do {
const int8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const int8_t*) ((uintptr_t) a0 + a_offset);
}
const int8_t* restrict a1 = a[1];
if XNN_UNPREDICTABLE(a1 != zero) {
a1 = (const int8_t*) ((uintptr_t) a1 + a_offset);
}
const int8_t* restrict a2 = a[2];
if XNN_UNPREDICTABLE(a2 != zero) {
a2 = (const int8_t*) ((uintptr_t) a2 + a_offset);
}
a += 3;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepi8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepi8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
p -= 3 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale01234567 = _mm256_load_ps(w);
w = (const void*) ((const float*) w + 8);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale01234567);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale01234567);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale01234567);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storel_epi64((__m128i*) c0, vout_lo);
c2 = (int8_t*) ((uintptr_t) c2 + cn_stride);
c1 = (int8_t*) ((uintptr_t) c1 + cn_stride);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a = (const int8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
_mm_storeu_si32(c1, vout_hi);
_mm_storeu_si32(c0, vout_lo);
c2 += 4;
c1 += 4;
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c2 += 2;
c1 += 2;
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c2 = (int8_t) _mm_extract_epi8(vout_lo, 8);
*c1 = (int8_t) _mm_extract_epi8(vout_hi, 0);
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qs8_dwconv_minmax_fp32_ukernel_up16x25__avx2_mul32(
size_t channels,
size_t output_width,
const int8_t** input,
const void* weights,
int8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const int8_t* zero,
const union xnn_qs8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
do {
const int8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const int8_t*) ((uintptr_t) i0 + input_offset);
}
const int8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const int8_t*) ((uintptr_t) i1 + input_offset);
}
const int8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const int8_t*) ((uintptr_t) i2 + input_offset);
}
const int8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const int8_t*) ((uintptr_t) i3 + input_offset);
}
const int8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const int8_t*) ((uintptr_t) i4 + input_offset);
}
const int8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const int8_t*) ((uintptr_t) i5 + input_offset);
}
const int8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const int8_t*) ((uintptr_t) i6 + input_offset);
}
const int8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const int8_t*) ((uintptr_t) i7 + input_offset);
}
const int8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const int8_t*) ((uintptr_t) i8 + input_offset);
}
const int8_t* i9 = input[9];
assert(i9 != NULL);
if XNN_UNPREDICTABLE(i9 != zero) {
i9 = (const int8_t*) ((uintptr_t) i9 + input_offset);
}
const int8_t* i10 = input[10];
assert(i10 != NULL);
if XNN_UNPREDICTABLE(i10 != zero) {
i10 = (const int8_t*) ((uintptr_t) i10 + input_offset);
}
const int8_t* i11 = input[11];
assert(i11 != NULL);
if XNN_UNPREDICTABLE(i11 != zero) {
i11 = (const int8_t*) ((uintptr_t) i11 + input_offset);
}
const int8_t* i12 = input[12];
assert(i12 != NULL);
if XNN_UNPREDICTABLE(i12 != zero) {
i12 = (const int8_t*) ((uintptr_t) i12 + input_offset);
}
const int8_t* i13 = input[13];
assert(i13 != NULL);
if XNN_UNPREDICTABLE(i13 != zero) {
i13 = (const int8_t*) ((uintptr_t) i13 + input_offset);
}
const int8_t* i14 = input[14];
assert(i14 != NULL);
if XNN_UNPREDICTABLE(i14 != zero) {
i14 = (const int8_t*) ((uintptr_t) i14 + input_offset);
}
const int8_t* i15 = input[15];
assert(i15 != NULL);
if XNN_UNPREDICTABLE(i15 != zero) {
i15 = (const int8_t*) ((uintptr_t) i15 + input_offset);
}
const int8_t* i16 = input[16];
assert(i16 != NULL);
if XNN_UNPREDICTABLE(i16 != zero) {
i16 = (const int8_t*) ((uintptr_t) i16 + input_offset);
}
const int8_t* i17 = input[17];
assert(i17 != NULL);
if XNN_UNPREDICTABLE(i17 != zero) {
i17 = (const int8_t*) ((uintptr_t) i17 + input_offset);
}
const int8_t* i18 = input[18];
assert(i18 != NULL);
if XNN_UNPREDICTABLE(i18 != zero) {
i18 = (const int8_t*) ((uintptr_t) i18 + input_offset);
}
const int8_t* i19 = input[19];
assert(i19 != NULL);
if XNN_UNPREDICTABLE(i19 != zero) {
i19 = (const int8_t*) ((uintptr_t) i19 + input_offset);
}
const int8_t* i20 = input[20];
assert(i20 != NULL);
if XNN_UNPREDICTABLE(i20 != zero) {
i20 = (const int8_t*) ((uintptr_t) i20 + input_offset);
}
const int8_t* i21 = input[21];
assert(i21 != NULL);
if XNN_UNPREDICTABLE(i21 != zero) {
i21 = (const int8_t*) ((uintptr_t) i21 + input_offset);
}
const int8_t* i22 = input[22];
assert(i22 != NULL);
if XNN_UNPREDICTABLE(i22 != zero) {
i22 = (const int8_t*) ((uintptr_t) i22 + input_offset);
}
const int8_t* i23 = input[23];
assert(i23 != NULL);
if XNN_UNPREDICTABLE(i23 != zero) {
i23 = (const int8_t*) ((uintptr_t) i23 + input_offset);
}
const int8_t* i24 = input[24];
assert(i24 != NULL);
if XNN_UNPREDICTABLE(i24 != zero) {
i24 = (const int8_t*) ((uintptr_t) i24 + input_offset);
}
input = (const int8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(int8_t))));
const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(int8_t))));
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(int8_t))));
const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(int8_t))));
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(int8_t))));
const __m256i vi2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(int8_t))));
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t))));
const __m256i vi3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(int8_t))));
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(int8_t))));
const __m256i vi4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(int8_t))));
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(int8_t))));
const __m256i vi5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(int8_t))));
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(int8_t))));
const __m256i vi6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(int8_t))));
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(int8_t))));
const __m256i vi7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(int8_t))));
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(int8_t))));
const __m256i vi8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(int8_t))));
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
const __m256i vi9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(int8_t))));
const __m256i vi9x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i9 + 8)));
const __m256i vk9x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 152 * sizeof(int8_t))));
i9 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi9x89ABCDEF, vk9x89ABCDEF));
const __m256i vi10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 160 * sizeof(int8_t))));
const __m256i vi10x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i10 + 8)));
const __m256i vk10x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 168 * sizeof(int8_t))));
i10 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi10x89ABCDEF, vk10x89ABCDEF));
const __m256i vi11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 176 * sizeof(int8_t))));
const __m256i vi11x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i11 + 8)));
const __m256i vk11x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 184 * sizeof(int8_t))));
i11 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi11x89ABCDEF, vk11x89ABCDEF));
const __m256i vi12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 192 * sizeof(int8_t))));
const __m256i vi12x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i12 + 8)));
const __m256i vk12x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 200 * sizeof(int8_t))));
i12 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi12x89ABCDEF, vk12x89ABCDEF));
const __m256i vi13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 208 * sizeof(int8_t))));
const __m256i vi13x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i13 + 8)));
const __m256i vk13x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 216 * sizeof(int8_t))));
i13 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi13x89ABCDEF, vk13x89ABCDEF));
const __m256i vi14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 224 * sizeof(int8_t))));
const __m256i vi14x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i14 + 8)));
const __m256i vk14x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 232 * sizeof(int8_t))));
i14 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi14x89ABCDEF, vk14x89ABCDEF));
const __m256i vi15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 240 * sizeof(int8_t))));
const __m256i vi15x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i15 + 8)));
const __m256i vk15x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 248 * sizeof(int8_t))));
i15 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi15x89ABCDEF, vk15x89ABCDEF));
const __m256i vi16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 256 * sizeof(int8_t))));
const __m256i vi16x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i16 + 8)));
const __m256i vk16x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 264 * sizeof(int8_t))));
i16 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi16x89ABCDEF, vk16x89ABCDEF));
const __m256i vi17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 272 * sizeof(int8_t))));
const __m256i vi17x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i17 + 8)));
const __m256i vk17x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 280 * sizeof(int8_t))));
i17 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi17x89ABCDEF, vk17x89ABCDEF));
const __m256i vi18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 288 * sizeof(int8_t))));
const __m256i vi18x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i18 + 8)));
const __m256i vk18x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 296 * sizeof(int8_t))));
i18 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi18x89ABCDEF, vk18x89ABCDEF));
const __m256i vi19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 304 * sizeof(int8_t))));
const __m256i vi19x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i19 + 8)));
const __m256i vk19x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 312 * sizeof(int8_t))));
i19 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi19x89ABCDEF, vk19x89ABCDEF));
const __m256i vi20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 320 * sizeof(int8_t))));
const __m256i vi20x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i20 + 8)));
const __m256i vk20x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 328 * sizeof(int8_t))));
i20 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi20x89ABCDEF, vk20x89ABCDEF));
const __m256i vi21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 336 * sizeof(int8_t))));
const __m256i vi21x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i21 + 8)));
const __m256i vk21x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 344 * sizeof(int8_t))));
i21 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi21x89ABCDEF, vk21x89ABCDEF));
const __m256i vi22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 352 * sizeof(int8_t))));
const __m256i vi22x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i22 + 8)));
const __m256i vk22x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 360 * sizeof(int8_t))));
i22 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi22x89ABCDEF, vk22x89ABCDEF));
const __m256i vi23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 368 * sizeof(int8_t))));
const __m256i vi23x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i23 + 8)));
const __m256i vk23x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 376 * sizeof(int8_t))));
i23 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi23x89ABCDEF, vk23x89ABCDEF));
const __m256i vi24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 384 * sizeof(int8_t))));
const __m256i vi24x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i24 + 8)));
const __m256i vk24x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 392 * sizeof(int8_t))));
i24 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi24x89ABCDEF, vk24x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 400 * sizeof(int8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const int8_t* k = (const int8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16)));
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32)));
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48)));
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64)));
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80)));
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96)));
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112)));
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128)));
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
const __m256i vi9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 144)));
i9 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
const __m256i vi10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 160)));
i10 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
const __m256i vi11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 176)));
i11 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
const __m256i vi12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 192)));
i12 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
const __m256i vi13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 208)));
i13 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
const __m256i vi14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 224)));
i14 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
const __m256i vi15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 240)));
i15 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
const __m256i vi16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 256)));
i16 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
const __m256i vi17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 272)));
i17 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
const __m256i vi18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 288)));
i18 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
const __m256i vi19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 304)));
i19 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
const __m256i vi20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 320)));
i20 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
const __m256i vi21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 336)));
i21 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
const __m256i vi22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 352)));
i22 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
const __m256i vi23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 368)));
i23 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
const __m256i vi24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 384)));
i24 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.scale));
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (int8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qs8_dwconv_minmax_fp32_ukernel_up16x9__avx2_mul32(
size_t channels,
size_t output_width,
const int8_t** input,
const void* weights,
int8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const int8_t* zero,
const union xnn_qs8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
do {
const int8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const int8_t*) ((uintptr_t) i0 + input_offset);
}
const int8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const int8_t*) ((uintptr_t) i1 + input_offset);
}
const int8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const int8_t*) ((uintptr_t) i2 + input_offset);
}
const int8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const int8_t*) ((uintptr_t) i3 + input_offset);
}
const int8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const int8_t*) ((uintptr_t) i4 + input_offset);
}
const int8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const int8_t*) ((uintptr_t) i5 + input_offset);
}
const int8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const int8_t*) ((uintptr_t) i6 + input_offset);
}
const int8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const int8_t*) ((uintptr_t) i7 + input_offset);
}
const int8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const int8_t*) ((uintptr_t) i8 + input_offset);
}
input = (const int8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(int8_t))));
const __m256i vi0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(int8_t))));
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(int8_t))));
const __m256i vi1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(int8_t))));
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(int8_t))));
const __m256i vi2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(int8_t))));
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(int8_t))));
const __m256i vi3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(int8_t))));
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(int8_t))));
const __m256i vi4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(int8_t))));
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(int8_t))));
const __m256i vi5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(int8_t))));
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(int8_t))));
const __m256i vi6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(int8_t))));
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(int8_t))));
const __m256i vi7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(int8_t))));
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(int8_t))));
const __m256i vi8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(int8_t))));
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(int8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const int8_t* k = (const int8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) k));
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16)));
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32)));
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48)));
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64)));
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80)));
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96)));
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112)));
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128)));
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.scale));
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (int8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qs8_f32_vcvt_ukernel__avx2_x16(
size_t n,
const int8_t* x,
float* y,
const union xnn_qs8_f32_cvt_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(int8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vminus_zero_point = _mm256_load_si256((const __m256i*) params->avx.minus_zero_point);
const __m256 vscale = _mm256_load_ps(params->avx.scale);
for (; n >= 16 * sizeof(int8_t); n -= 16 * sizeof(int8_t)) {
__m256i vx01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) x));
__m256i vx89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (x + 8)));
x += 16;
vx01234567 = _mm256_add_epi32(vx01234567, vminus_zero_point);
vx89ABCDEF = _mm256_add_epi32(vx89ABCDEF, vminus_zero_point);
__m256 vy01234567 = _mm256_cvtepi32_ps(vx01234567);
__m256 vy89ABCDEF = _mm256_cvtepi32_ps(vx89ABCDEF);
vy01234567 = _mm256_mul_ps(vy01234567, vscale);
vy89ABCDEF = _mm256_mul_ps(vy89ABCDEF, vscale);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
y += 16;
}
for (; n >= 8 * sizeof(int8_t); n -= 8 * sizeof(int8_t)) {
__m256i vx = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) x));
vx = _mm256_add_epi32(vx, vminus_zero_point);
x += 8;
__m256 vy = _mm256_cvtepi32_ps(vx);
vy = _mm256_mul_ps(vy, vscale);
_mm256_storeu_ps(y, vy);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(int8_t));
assert(n <= 7 * sizeof(int8_t));
__m256i vx = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) x));
vx = _mm256_add_epi32(vx, vminus_zero_point);
__m256 vy = _mm256_cvtepi32_ps(vx);
vy = _mm256_mul_ps(vy, vscale);
__m128 vy_lo = _mm256_castps256_ps128(vy);
if (n & (4 * sizeof(int8_t))) {
_mm_storeu_ps(y, vy_lo);
vy_lo = _mm256_extractf128_ps(vy, 1);
y += 4;
}
if (n & (2 * sizeof(int8_t))) {
_mm_storel_pi((__m64*) y, vy_lo);
vy_lo = _mm_movehl_ps(vy_lo, vy_lo);
y += 2;
}
if (n & (1 * sizeof(int8_t))) {
_mm_store_ss(y, vy_lo);
}
}
}
void xnn_qs8_gemm_minmax_fp32_ukernel_1x8c8__avx2(
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_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
const int8_t* a0 = a;
int8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a0 = (const int8_t*) ((uintptr_t) a0 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qs8_gemm_minmax_fp32_ukernel_3x8c8__avx2(
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_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
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;
}
const int8_t* a2 = (const int8_t*) ((uintptr_t) a1 + a_stride);
int8_t* c2 = (int8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
a2 = a1;
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepi8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepi8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
c1 = (int8_t*) ((uintptr_t) c1 + cn_stride);
c2 = (int8_t*) ((uintptr_t) c2 + cn_stride);
a0 = (const int8_t*) ((uintptr_t) a0 - kc);
a1 = (const int8_t*) ((uintptr_t) a1 - kc);
a2 = (const int8_t*) ((uintptr_t) a2 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
_mm_storeu_si32(c1, vout_hi);
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
c0 += 4;
c1 += 4;
c2 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
c0 += 2;
c1 += 2;
c2 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
*c1 = (int8_t) _mm_extract_epi8(vout_hi, 0);
*c2 = (int8_t) _mm_extract_epi8(vout_lo, 8);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qs8_igemm_minmax_fp32_ukernel_1x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const int8_t** restrict a,
const void* restrict w,
int8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const int8_t* zero,
const union xnn_qs8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (1 * sizeof(void*)) == 0);
assert(a_offset % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
int8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t p = ks;
do {
const int8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const int8_t*) ((uintptr_t) a0 + a_offset);
}
a += 1;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
p -= 1 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a = (const int8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qs8_igemm_minmax_fp32_ukernel_3x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const int8_t** restrict a,
const void* restrict w,
int8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const int8_t* zero,
const union xnn_qs8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (3 * sizeof(void*)) == 0);
assert(a_offset % sizeof(int8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
int8_t* c0 = c;
int8_t* c1 = (int8_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
c1 = c0;
}
int8_t* c2 = (int8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t p = ks;
do {
const int8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const int8_t*) ((uintptr_t) a0 + a_offset);
}
const int8_t* restrict a1 = a[1];
if XNN_UNPREDICTABLE(a1 != zero) {
a1 = (const int8_t*) ((uintptr_t) a1 + a_offset);
}
const int8_t* restrict a2 = a[2];
if XNN_UNPREDICTABLE(a2 != zero) {
a2 = (const int8_t*) ((uintptr_t) a2 + a_offset);
}
a += 3;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepi8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepi8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepi8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_cvtepi8_epi16(vb01);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 16));
const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 32));
const __m256i vxb45 = _mm256_cvtepi8_epi16(vb45);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const int8_t*) w + 48));
const __m256i vxb67 = _mm256_cvtepi8_epi16(vb67);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const int8_t*) w + 64);
k += 8 * sizeof(int8_t);
}
p -= 3 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packs_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epi8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storel_epi64((__m128i*) c0, vout_lo);
c2 = (int8_t*) ((uintptr_t) c2 + cn_stride);
c1 = (int8_t*) ((uintptr_t) c1 + cn_stride);
c0 = (int8_t*) ((uintptr_t) c0 + cn_stride);
a = (const int8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
_mm_storeu_si32(c1, vout_hi);
_mm_storeu_si32(c0, vout_lo);
c2 += 4;
c1 += 4;
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c2 += 2;
c1 += 2;
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c2 = (int8_t) _mm_extract_epi8(vout_lo, 8);
*c1 = (int8_t) _mm_extract_epi8(vout_hi, 0);
*c0 = (int8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16(
size_t n,
const int8_t* input_a,
const int8_t* input_b,
int8_t* output,
const union xnn_qs8_add_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
const __m256i vbias = _mm256_load_si256((const __m256i*) params->avx2.bias);
const __m256i va_multiplier = _mm256_load_si256((const __m256i*) params->avx2.a_multiplier);
const __m256i vb_multiplier = _mm256_load_si256((const __m256i*) params->avx2.b_multiplier);
const __m128i vshift = _mm_load_si128((const __m128i*) params->avx2.shift);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->avx2.output_min);
const __m128i voutput_max = _mm_load_si128((const __m128i*) params->avx2.output_max);
for (; n >= 16 * sizeof(int8_t); n -= 16 * sizeof(int8_t)) {
const __m256i va01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i vb01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_b));
const __m256i va89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_a + 8)));
const __m256i vb89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_b + 8)));
input_a += 16;
input_b += 16;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
__m256i vacc89ABCDEF = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va89ABCDEF, va_multiplier));
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vb01234567, vb_multiplier));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vb89ABCDEF, vb_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
vacc89ABCDEF = _mm256_sra_epi32(vacc89ABCDEF, vshift);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
vout0123456789ABCDEF = _mm_min_epi8(vout0123456789ABCDEF, voutput_max);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(n != 0) {
do {
const __m256i va01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i vb01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_b));
input_a += 8;
input_b += 8;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vb01234567, vb_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), _mm256_castsi256_si128(voutput_zero_point));
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
vout0123456701234567 = _mm_min_epi8(vout0123456701234567, voutput_max);
if XNN_LIKELY(n >= (8 * sizeof(int8_t))) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
n -= 8 * sizeof(int8_t);
} else {
if (n & (4 * sizeof(int8_t))) {
_mm_storeu_si32(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (n & (2 * sizeof(int8_t))) {
_mm_storeu_si16(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (n & (1 * sizeof(int8_t))) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
}
n = 0;
}
} while (n != 0);
}
}
void xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16(
size_t n,
const int8_t* input_a,
const int8_t* input_b,
int8_t* output,
const union xnn_qs8_add_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
const __m256i va_multiplier = _mm256_load_si256((const __m256i*) params->avx2.a_multiplier);
const __m128i vshift = _mm_load_si128((const __m128i*) params->avx2.shift);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->avx2.output_min);
const __m128i voutput_max = _mm_load_si128((const __m128i*) params->avx2.output_max);
const __m256i vbias = _mm256_add_epi32(
_mm256_broadcastd_epi32(_mm_cvtsi32_si128(params->avx2.b_multiplier[0] * (int32_t) *input_b)),
_mm256_load_si256((const __m256i*) params->avx2.bias));
for (; n >= 16 * sizeof(int8_t); n -= 16 * sizeof(int8_t)) {
const __m256i va01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i va89ABCDEF = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) (input_a + 8)));
input_a += 16;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
__m256i vacc89ABCDEF = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va89ABCDEF, va_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
vacc89ABCDEF = _mm256_sra_epi32(vacc89ABCDEF, vshift);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packs_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
vout0123456789ABCDEF = _mm_max_epi8(vout0123456789ABCDEF, voutput_min);
vout0123456789ABCDEF = _mm_min_epi8(vout0123456789ABCDEF, voutput_max);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(n != 0) {
do {
const __m256i va01234567 = _mm256_cvtepi8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
input_a += 8;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), _mm256_castsi256_si128(voutput_zero_point));
__m128i vout0123456701234567 = _mm_packs_epi16(vout01234567, vout01234567);
vout0123456701234567 = _mm_max_epi8(vout0123456701234567, voutput_min);
vout0123456701234567 = _mm_min_epi8(vout0123456701234567, voutput_max);
if XNN_LIKELY(n >= (8 * sizeof(int8_t))) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
n -= 8 * sizeof(int8_t);
} else {
if (n & (4 * sizeof(int8_t))) {
_mm_storeu_si32(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (n & (2 * sizeof(int8_t))) {
_mm_storeu_si16(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (n & (1 * sizeof(int8_t))) {
*output = (int8_t) _mm_extract_epi8(vout0123456701234567, 0);
}
n = 0;
}
} while (n != 0);
}
}
void xnn_qs8_vcvt_ukernel__avx2_x32(
size_t n,
const int8_t* x,
int8_t* y,
const union xnn_qs8_cvt_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(int8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vinput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.input_zero_point);
const __m256i vmultiplier = _mm256_load_si256((const __m256i*) params->avx2.multiplier);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
for (; n >= 32 * sizeof(int8_t); n -= 32 * sizeof(int8_t)) {
__m256i vacc0 = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vacc1 = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) (x + 16)));
x += 32;
vacc0 = _mm256_sub_epi16(vinput_zero_point, vacc0);
vacc1 = _mm256_sub_epi16(vinput_zero_point, vacc1);
vacc0 = _mm256_slli_epi16(vacc0, 7);
vacc1 = _mm256_slli_epi16(vacc1, 7);
vacc0 = _mm256_mulhrs_epi16(vacc0, vmultiplier);
vacc1 = _mm256_mulhrs_epi16(vacc1, vmultiplier);
vacc0 = _mm256_adds_epi16(vacc0, voutput_zero_point);
vacc1 = _mm256_adds_epi16(vacc1, voutput_zero_point);
__m256i vy0 = _mm256_packs_epi16(vacc0, vacc1);
vy0 = _mm256_permute4x64_epi64(vy0, _MM_SHUFFLE(3, 1, 2, 0));
_mm256_storeu_si256((__m256i*) y, vy0);
y += 32;
}
for (; n >= 16 * sizeof(int8_t); n -= 16 * sizeof(int8_t)) {
__m256i vacc = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
x += 16;
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
const __m128i vy = _mm_packs_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
_mm_storeu_si128((__m128i*) y, vy);
y += 16;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(int8_t));
assert(n <= 15 * sizeof(int8_t));
__m256i vacc = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
__m128i vy = _mm_packs_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
if (n & (8 * sizeof(int8_t))) {
_mm_storel_epi64((__m128i*) y, vy);
vy = _mm_unpackhi_epi64(vy, vy);
y += 8;
}
if (n & (4 * sizeof(int8_t))) {
_mm_storeu_si32(y, vy);
vy = _mm_srli_epi64(vy, 32);
y += 4;
}
if (n & (2 * sizeof(int8_t))) {
_mm_storeu_si16(y, vy);
vy = _mm_srli_epi32(vy, 16);
y += 2;
}
if (n & (1 * sizeof(int8_t))) {
*y = (int8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_qs8_vlrelu_ukernel__avx2_x32(
size_t n,
const int8_t* x,
int8_t* y,
const union xnn_qs8_lrelu_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(int8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vinput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.input_zero_point);
const __m256i vpositive_multiplier = _mm256_load_si256((const __m256i*) params->avx2.positive_multiplier);
const __m256i vnegative_multiplier = _mm256_load_si256((const __m256i*) params->avx2.negative_multiplier);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
for (; n >= 32 * sizeof(int8_t); n -= 32 * sizeof(int8_t)) {
__m256i vacc0 = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vacc1 = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) (x + 16)));
x += 32;
__m256i vmultiplier0 = _mm256_cmpgt_epi16(vacc0, vinput_zero_point);
vacc0 = _mm256_sub_epi16(vinput_zero_point, vacc0);
__m256i vmultiplier1 = _mm256_cmpgt_epi16(vacc1, vinput_zero_point);
vacc1 = _mm256_sub_epi16(vinput_zero_point, vacc1);
vmultiplier0 = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier0);
vacc0 = _mm256_slli_epi16(vacc0, 7);
vmultiplier1 = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier1);
vacc1 = _mm256_slli_epi16(vacc1, 7);
vacc0 = _mm256_mulhrs_epi16(vacc0, vmultiplier0);
vacc1 = _mm256_mulhrs_epi16(vacc1, vmultiplier1);
vacc0 = _mm256_adds_epi16(vacc0, voutput_zero_point);
vacc1 = _mm256_adds_epi16(vacc1, voutput_zero_point);
__m256i vy0 = _mm256_packs_epi16(vacc0, vacc1);
vy0 = _mm256_permute4x64_epi64(vy0, _MM_SHUFFLE(3, 1, 2, 0));
_mm256_storeu_si256((__m256i*) y, vy0);
y += 32;
}
for (; n >= 16 * sizeof(int8_t); n -= 16 * sizeof(int8_t)) {
__m256i vacc = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vmultiplier = _mm256_cmpgt_epi16(vacc, vinput_zero_point);
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vmultiplier = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
x += 16;
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
const __m128i vy = _mm_packs_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
_mm_storeu_si128((__m128i*) y, vy);
y += 16;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(int8_t));
assert(n <= 15 * sizeof(int8_t));
__m256i vacc = _mm256_cvtepi8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vmultiplier = _mm256_cmpgt_epi16(vacc, vinput_zero_point);
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vmultiplier = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
__m128i vy = _mm_packs_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
if (n & (8 * sizeof(int8_t))) {
_mm_storel_epi64((__m128i*) y, vy);
vy = _mm_unpackhi_epi64(vy, vy);
y += 8;
}
if (n & (4 * sizeof(int8_t))) {
_mm_storeu_si32(y, vy);
vy = _mm_srli_epi64(vy, 32);
y += 4;
}
if (n & (2 * sizeof(int8_t))) {
_mm_storeu_si16(y, vy);
vy = _mm_srli_epi32(vy, 16);
y += 2;
}
if (n & (1 * sizeof(int8_t))) {
*y = (int8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_qu8_dwconv_minmax_fp32_ukernel_up16x25__avx2_mul32(
size_t channels,
size_t output_width,
const uint8_t** input,
const void* weights,
uint8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const uint8_t* zero,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
const __m256i vk_zero_point = _mm256_cvtepu16_epi32(_mm_load_si128((const __m128i*) params->fp32_avx2.kernel_zero_point));
do {
const uint8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint8_t*) ((uintptr_t) i0 + input_offset);
}
const uint8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint8_t*) ((uintptr_t) i1 + input_offset);
}
const uint8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint8_t*) ((uintptr_t) i2 + input_offset);
}
const uint8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint8_t*) ((uintptr_t) i3 + input_offset);
}
const uint8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint8_t*) ((uintptr_t) i4 + input_offset);
}
const uint8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint8_t*) ((uintptr_t) i5 + input_offset);
}
const uint8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint8_t*) ((uintptr_t) i6 + input_offset);
}
const uint8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint8_t*) ((uintptr_t) i7 + input_offset);
}
const uint8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const uint8_t*) ((uintptr_t) i8 + input_offset);
}
const uint8_t* i9 = input[9];
assert(i9 != NULL);
if XNN_UNPREDICTABLE(i9 != zero) {
i9 = (const uint8_t*) ((uintptr_t) i9 + input_offset);
}
const uint8_t* i10 = input[10];
assert(i10 != NULL);
if XNN_UNPREDICTABLE(i10 != zero) {
i10 = (const uint8_t*) ((uintptr_t) i10 + input_offset);
}
const uint8_t* i11 = input[11];
assert(i11 != NULL);
if XNN_UNPREDICTABLE(i11 != zero) {
i11 = (const uint8_t*) ((uintptr_t) i11 + input_offset);
}
const uint8_t* i12 = input[12];
assert(i12 != NULL);
if XNN_UNPREDICTABLE(i12 != zero) {
i12 = (const uint8_t*) ((uintptr_t) i12 + input_offset);
}
const uint8_t* i13 = input[13];
assert(i13 != NULL);
if XNN_UNPREDICTABLE(i13 != zero) {
i13 = (const uint8_t*) ((uintptr_t) i13 + input_offset);
}
const uint8_t* i14 = input[14];
assert(i14 != NULL);
if XNN_UNPREDICTABLE(i14 != zero) {
i14 = (const uint8_t*) ((uintptr_t) i14 + input_offset);
}
const uint8_t* i15 = input[15];
assert(i15 != NULL);
if XNN_UNPREDICTABLE(i15 != zero) {
i15 = (const uint8_t*) ((uintptr_t) i15 + input_offset);
}
const uint8_t* i16 = input[16];
assert(i16 != NULL);
if XNN_UNPREDICTABLE(i16 != zero) {
i16 = (const uint8_t*) ((uintptr_t) i16 + input_offset);
}
const uint8_t* i17 = input[17];
assert(i17 != NULL);
if XNN_UNPREDICTABLE(i17 != zero) {
i17 = (const uint8_t*) ((uintptr_t) i17 + input_offset);
}
const uint8_t* i18 = input[18];
assert(i18 != NULL);
if XNN_UNPREDICTABLE(i18 != zero) {
i18 = (const uint8_t*) ((uintptr_t) i18 + input_offset);
}
const uint8_t* i19 = input[19];
assert(i19 != NULL);
if XNN_UNPREDICTABLE(i19 != zero) {
i19 = (const uint8_t*) ((uintptr_t) i19 + input_offset);
}
const uint8_t* i20 = input[20];
assert(i20 != NULL);
if XNN_UNPREDICTABLE(i20 != zero) {
i20 = (const uint8_t*) ((uintptr_t) i20 + input_offset);
}
const uint8_t* i21 = input[21];
assert(i21 != NULL);
if XNN_UNPREDICTABLE(i21 != zero) {
i21 = (const uint8_t*) ((uintptr_t) i21 + input_offset);
}
const uint8_t* i22 = input[22];
assert(i22 != NULL);
if XNN_UNPREDICTABLE(i22 != zero) {
i22 = (const uint8_t*) ((uintptr_t) i22 + input_offset);
}
const uint8_t* i23 = input[23];
assert(i23 != NULL);
if XNN_UNPREDICTABLE(i23 != zero) {
i23 = (const uint8_t*) ((uintptr_t) i23 + input_offset);
}
const uint8_t* i24 = input[24];
assert(i24 != NULL);
if XNN_UNPREDICTABLE(i24 != zero) {
i24 = (const uint8_t*) ((uintptr_t) i24 + input_offset);
}
input = (const uint8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi0x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(uint8_t)))), vk_zero_point);
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi1x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(uint8_t)))), vk_zero_point);
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi2x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(uint8_t)))), vk_zero_point);
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi3x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(uint8_t)))), vk_zero_point);
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi4x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(uint8_t)))), vk_zero_point);
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi5x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(uint8_t)))), vk_zero_point);
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi6x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(uint8_t)))), vk_zero_point);
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi7x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(uint8_t)))), vk_zero_point);
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi8x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(uint8_t)))), vk_zero_point);
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
const __m256i vi9x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi9x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i9 + 8)));
const __m256i vk9x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 152 * sizeof(uint8_t)))), vk_zero_point);
i9 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi9x89ABCDEF, vk9x89ABCDEF));
const __m256i vi10x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 160 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi10x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i10 + 8)));
const __m256i vk10x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 168 * sizeof(uint8_t)))), vk_zero_point);
i10 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi10x89ABCDEF, vk10x89ABCDEF));
const __m256i vi11x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 176 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi11x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i11 + 8)));
const __m256i vk11x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 184 * sizeof(uint8_t)))), vk_zero_point);
i11 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi11x89ABCDEF, vk11x89ABCDEF));
const __m256i vi12x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 192 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi12x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i12 + 8)));
const __m256i vk12x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 200 * sizeof(uint8_t)))), vk_zero_point);
i12 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi12x89ABCDEF, vk12x89ABCDEF));
const __m256i vi13x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 208 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi13x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i13 + 8)));
const __m256i vk13x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 216 * sizeof(uint8_t)))), vk_zero_point);
i13 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi13x89ABCDEF, vk13x89ABCDEF));
const __m256i vi14x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 224 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi14x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i14 + 8)));
const __m256i vk14x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 232 * sizeof(uint8_t)))), vk_zero_point);
i14 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi14x89ABCDEF, vk14x89ABCDEF));
const __m256i vi15x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 240 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi15x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i15 + 8)));
const __m256i vk15x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 248 * sizeof(uint8_t)))), vk_zero_point);
i15 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi15x89ABCDEF, vk15x89ABCDEF));
const __m256i vi16x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 256 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi16x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i16 + 8)));
const __m256i vk16x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 264 * sizeof(uint8_t)))), vk_zero_point);
i16 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi16x89ABCDEF, vk16x89ABCDEF));
const __m256i vi17x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 272 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi17x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i17 + 8)));
const __m256i vk17x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 280 * sizeof(uint8_t)))), vk_zero_point);
i17 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi17x89ABCDEF, vk17x89ABCDEF));
const __m256i vi18x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 288 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi18x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i18 + 8)));
const __m256i vk18x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 296 * sizeof(uint8_t)))), vk_zero_point);
i18 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi18x89ABCDEF, vk18x89ABCDEF));
const __m256i vi19x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 304 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi19x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i19 + 8)));
const __m256i vk19x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 312 * sizeof(uint8_t)))), vk_zero_point);
i19 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi19x89ABCDEF, vk19x89ABCDEF));
const __m256i vi20x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 320 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi20x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i20 + 8)));
const __m256i vk20x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 328 * sizeof(uint8_t)))), vk_zero_point);
i20 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi20x89ABCDEF, vk20x89ABCDEF));
const __m256i vi21x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 336 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi21x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i21 + 8)));
const __m256i vk21x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 344 * sizeof(uint8_t)))), vk_zero_point);
i21 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi21x89ABCDEF, vk21x89ABCDEF));
const __m256i vi22x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 352 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi22x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i22 + 8)));
const __m256i vk22x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 360 * sizeof(uint8_t)))), vk_zero_point);
i22 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi22x89ABCDEF, vk22x89ABCDEF));
const __m256i vi23x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 368 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi23x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i23 + 8)));
const __m256i vk23x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 376 * sizeof(uint8_t)))), vk_zero_point);
i23 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi23x89ABCDEF, vk23x89ABCDEF));
const __m256i vi24x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 384 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi24x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i24 + 8)));
const __m256i vk24x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 392 * sizeof(uint8_t)))), vk_zero_point);
i24 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi24x89ABCDEF, vk24x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 400 * sizeof(uint8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packus_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epu8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const uint8_t* k = (const uint8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) k)), vk_zero_point);
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16))), vk_zero_point);
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32))), vk_zero_point);
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48))), vk_zero_point);
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64))), vk_zero_point);
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80))), vk_zero_point);
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96))), vk_zero_point);
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112))), vk_zero_point);
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128))), vk_zero_point);
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
const __m256i vi9x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i9));
const __m256i vk9x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 144))), vk_zero_point);
i9 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi9x01234567, vk9x01234567));
const __m256i vi10x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i10));
const __m256i vk10x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 160))), vk_zero_point);
i10 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi10x01234567, vk10x01234567));
const __m256i vi11x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i11));
const __m256i vk11x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 176))), vk_zero_point);
i11 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi11x01234567, vk11x01234567));
const __m256i vi12x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i12));
const __m256i vk12x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 192))), vk_zero_point);
i12 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi12x01234567, vk12x01234567));
const __m256i vi13x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i13));
const __m256i vk13x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 208))), vk_zero_point);
i13 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi13x01234567, vk13x01234567));
const __m256i vi14x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i14));
const __m256i vk14x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 224))), vk_zero_point);
i14 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi14x01234567, vk14x01234567));
const __m256i vi15x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i15));
const __m256i vk15x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 240))), vk_zero_point);
i15 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi15x01234567, vk15x01234567));
const __m256i vi16x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i16));
const __m256i vk16x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 256))), vk_zero_point);
i16 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi16x01234567, vk16x01234567));
const __m256i vi17x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i17));
const __m256i vk17x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 272))), vk_zero_point);
i17 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi17x01234567, vk17x01234567));
const __m256i vi18x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i18));
const __m256i vk18x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 288))), vk_zero_point);
i18 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi18x01234567, vk18x01234567));
const __m256i vi19x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i19));
const __m256i vk19x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 304))), vk_zero_point);
i19 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi19x01234567, vk19x01234567));
const __m256i vi20x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i20));
const __m256i vk20x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 320))), vk_zero_point);
i20 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi20x01234567, vk20x01234567));
const __m256i vi21x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i21));
const __m256i vk21x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 336))), vk_zero_point);
i21 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi21x01234567, vk21x01234567));
const __m256i vi22x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i22));
const __m256i vk22x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 352))), vk_zero_point);
i22 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi22x01234567, vk22x01234567));
const __m256i vi23x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i23));
const __m256i vk23x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 368))), vk_zero_point);
i23 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi23x01234567, vk23x01234567));
const __m256i vi24x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i24));
const __m256i vk24x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 384))), vk_zero_point);
i24 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi24x01234567, vk24x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.scale));
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packus_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epu8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (uint8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (uint8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qu8_dwconv_minmax_fp32_ukernel_up16x9__avx2_mul32(
size_t channels,
size_t output_width,
const uint8_t** input,
const void* weights,
uint8_t* output,
size_t input_stride,
size_t output_increment,
size_t input_offset,
const uint8_t* zero,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(channels != 0);
assert(output_width != 0);
const __m256i vk_zero_point = _mm256_cvtepu16_epi32(_mm_load_si128((const __m128i*) params->fp32_avx2.kernel_zero_point));
do {
const uint8_t* i0 = input[0];
assert(i0 != NULL);
if XNN_UNPREDICTABLE(i0 != zero) {
i0 = (const uint8_t*) ((uintptr_t) i0 + input_offset);
}
const uint8_t* i1 = input[1];
assert(i1 != NULL);
if XNN_UNPREDICTABLE(i1 != zero) {
i1 = (const uint8_t*) ((uintptr_t) i1 + input_offset);
}
const uint8_t* i2 = input[2];
assert(i2 != NULL);
if XNN_UNPREDICTABLE(i2 != zero) {
i2 = (const uint8_t*) ((uintptr_t) i2 + input_offset);
}
const uint8_t* i3 = input[3];
assert(i3 != NULL);
if XNN_UNPREDICTABLE(i3 != zero) {
i3 = (const uint8_t*) ((uintptr_t) i3 + input_offset);
}
const uint8_t* i4 = input[4];
assert(i4 != NULL);
if XNN_UNPREDICTABLE(i4 != zero) {
i4 = (const uint8_t*) ((uintptr_t) i4 + input_offset);
}
const uint8_t* i5 = input[5];
assert(i5 != NULL);
if XNN_UNPREDICTABLE(i5 != zero) {
i5 = (const uint8_t*) ((uintptr_t) i5 + input_offset);
}
const uint8_t* i6 = input[6];
assert(i6 != NULL);
if XNN_UNPREDICTABLE(i6 != zero) {
i6 = (const uint8_t*) ((uintptr_t) i6 + input_offset);
}
const uint8_t* i7 = input[7];
assert(i7 != NULL);
if XNN_UNPREDICTABLE(i7 != zero) {
i7 = (const uint8_t*) ((uintptr_t) i7 + input_offset);
}
const uint8_t* i8 = input[8];
assert(i8 != NULL);
if XNN_UNPREDICTABLE(i8 != zero) {
i8 = (const uint8_t*) ((uintptr_t) i8 + input_offset);
}
input = (const uint8_t**) ((uintptr_t) input + input_stride);
size_t c = channels;
const void* w = weights;
for (; c >= 16; c -= 16) {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
__m256i vacc89ABCDEF = _mm256_loadu_si256((const __m256i*) ((const int32_t*) w + 8));
const __m256i vi0x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 0 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi0x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i0 + 8)));
const __m256i vk0x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 8 * sizeof(uint8_t)))), vk_zero_point);
i0 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi0x89ABCDEF, vk0x89ABCDEF));
const __m256i vi1x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 16 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi1x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i1 + 8)));
const __m256i vk1x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 24 * sizeof(uint8_t)))), vk_zero_point);
i1 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi1x89ABCDEF, vk1x89ABCDEF));
const __m256i vi2x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 32 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi2x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i2 + 8)));
const __m256i vk2x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 40 * sizeof(uint8_t)))), vk_zero_point);
i2 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi2x89ABCDEF, vk2x89ABCDEF));
const __m256i vi3x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 48 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi3x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i3 + 8)));
const __m256i vk3x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 56 * sizeof(uint8_t)))), vk_zero_point);
i3 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi3x89ABCDEF, vk3x89ABCDEF));
const __m256i vi4x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 64 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi4x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i4 + 8)));
const __m256i vk4x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 72 * sizeof(uint8_t)))), vk_zero_point);
i4 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi4x89ABCDEF, vk4x89ABCDEF));
const __m256i vi5x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 80 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi5x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i5 + 8)));
const __m256i vk5x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 88 * sizeof(uint8_t)))), vk_zero_point);
i5 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi5x89ABCDEF, vk5x89ABCDEF));
const __m256i vi6x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 96 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi6x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i6 + 8)));
const __m256i vk6x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 104 * sizeof(uint8_t)))), vk_zero_point);
i6 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi6x89ABCDEF, vk6x89ABCDEF));
const __m256i vi7x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 112 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi7x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i7 + 8)));
const __m256i vk7x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 120 * sizeof(uint8_t)))), vk_zero_point);
i7 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi7x89ABCDEF, vk7x89ABCDEF));
const __m256i vi8x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 128 * sizeof(uint8_t)))), vk_zero_point);
const __m256i vi8x89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (i8 + 8)));
const __m256i vk8x89ABCDEF = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) ((uintptr_t) w + 16 * sizeof(int32_t) + 136 * sizeof(uint8_t)))), vk_zero_point);
i8 += 16;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vi8x89ABCDEF, vk8x89ABCDEF));
w = (const void*) ((uintptr_t) w + 16 * sizeof(int32_t) + 144 * sizeof(uint8_t));
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
__m256 vscaled89ABCDEF = _mm256_cvtepi32_ps(vacc89ABCDEF);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, vscale);
vscaled89ABCDEF = _mm256_mul_ps(vscaled89ABCDEF, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled01234567 = _mm256_min_ps(vscaled01234567, voutput_max_less_zero_point);
vscaled89ABCDEF = _mm256_min_ps(vscaled89ABCDEF, voutput_max_less_zero_point);
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
vacc89ABCDEF = _mm256_cvtps_epi32(vscaled89ABCDEF);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packus_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456789ABCDEF = _mm_max_epu8(vout0123456789ABCDEF, voutput_min);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(c != 0) {
const uint8_t* k = (const uint8_t*) ((const int32_t*) w + 16);
do {
__m256i vacc01234567 = _mm256_loadu_si256((const __m256i*) w);
const __m256i vi0x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i0));
const __m256i vk0x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) k)), vk_zero_point);
i0 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi0x01234567, vk0x01234567));
const __m256i vi1x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i1));
const __m256i vk1x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 16))), vk_zero_point);
i1 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi1x01234567, vk1x01234567));
const __m256i vi2x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i2));
const __m256i vk2x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 32))), vk_zero_point);
i2 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi2x01234567, vk2x01234567));
const __m256i vi3x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i3));
const __m256i vk3x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 48))), vk_zero_point);
i3 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi3x01234567, vk3x01234567));
const __m256i vi4x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i4));
const __m256i vk4x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 64))), vk_zero_point);
i4 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi4x01234567, vk4x01234567));
const __m256i vi5x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i5));
const __m256i vk5x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 80))), vk_zero_point);
i5 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi5x01234567, vk5x01234567));
const __m256i vi6x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i6));
const __m256i vk6x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 96))), vk_zero_point);
i6 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi6x01234567, vk6x01234567));
const __m256i vi7x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i7));
const __m256i vk7x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 112))), vk_zero_point);
i7 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi7x01234567, vk7x01234567));
const __m256i vi8x01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) i8));
const __m256i vk8x01234567 = _mm256_sub_epi32(_mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (k + 128))), vk_zero_point);
i8 += 8;
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vi8x01234567, vk8x01234567));
k += 8;
__m256 vscaled01234567 = _mm256_cvtepi32_ps(vacc01234567);
vscaled01234567 = _mm256_mul_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.scale));
vscaled01234567 = _mm256_min_ps(vscaled01234567, _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point));
vacc01234567 = _mm256_cvtps_epi32(vscaled01234567);
w = (const void*) ((const int32_t*) w + 8);
const __m128i voutput_zero_point = _mm_load_si128((const __m128i*) params->fp32_avx2.output_zero_point);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), voutput_zero_point);
__m128i vout0123456701234567 = _mm_packus_epi16(vout01234567, vout01234567);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->fp32_avx2.output_min);
vout0123456701234567 = _mm_max_epu8(vout0123456701234567, voutput_min);
if XNN_LIKELY(c >= 8) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
c -= 8;
} else {
if (c & 4) {
unaligned_store_u32(output, (uint32_t) _mm_cvtsi128_si32(vout0123456701234567));
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (c & 2) {
unaligned_store_u16(output, (uint16_t) _mm_extract_epi16(vout0123456701234567, 0));
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (c & 1) {
*output = (uint8_t) _mm_extract_epi8(vout0123456701234567, 0);
output += 1;
}
c = 0;
}
} while (c != 0);
}
output = (uint8_t*) ((uintptr_t) output + output_increment);
} while (--output_width != 0);
}
void xnn_qu8_f32_vcvt_ukernel__avx2_x16(
size_t n,
const uint8_t* x,
float* y,
const union xnn_qu8_f32_cvt_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(uint8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vminus_zero_point = _mm256_load_si256((const __m256i*) params->avx.minus_zero_point);
const __m256 vscale = _mm256_load_ps(params->avx.scale);
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
__m256i vx01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) x));
__m256i vx89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (x + 8)));
x += 16;
vx01234567 = _mm256_add_epi32(vx01234567, vminus_zero_point);
vx89ABCDEF = _mm256_add_epi32(vx89ABCDEF, vminus_zero_point);
__m256 vy01234567 = _mm256_cvtepi32_ps(vx01234567);
__m256 vy89ABCDEF = _mm256_cvtepi32_ps(vx89ABCDEF);
vy01234567 = _mm256_mul_ps(vy01234567, vscale);
vy89ABCDEF = _mm256_mul_ps(vy89ABCDEF, vscale);
_mm256_storeu_ps(y, vy01234567);
_mm256_storeu_ps(y + 8, vy89ABCDEF);
y += 16;
}
for (; n >= 8 * sizeof(uint8_t); n -= 8 * sizeof(uint8_t)) {
__m256i vx = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) x));
vx = _mm256_add_epi32(vx, vminus_zero_point);
x += 8;
__m256 vy = _mm256_cvtepi32_ps(vx);
vy = _mm256_mul_ps(vy, vscale);
_mm256_storeu_ps(y, vy);
y += 8;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(uint8_t));
assert(n <= 7 * sizeof(uint8_t));
__m256i vx = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) x));
vx = _mm256_add_epi32(vx, vminus_zero_point);
__m256 vy = _mm256_cvtepi32_ps(vx);
vy = _mm256_mul_ps(vy, vscale);
__m128 vy_lo = _mm256_castps256_ps128(vy);
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_ps(y, vy_lo);
vy_lo = _mm256_extractf128_ps(vy, 1);
y += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storel_pi((__m64*) y, vy_lo);
vy_lo = _mm_movehl_ps(vy_lo, vy_lo);
y += 2;
}
if (n & (1 * sizeof(uint8_t))) {
_mm_store_ss(y, vy_lo);
}
}
}
void xnn_qu8_gemm_minmax_fp32_ukernel_1x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
const uint8_t* restrict a,
size_t a_stride,
const void* restrict w,
uint8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
const uint8_t* a0 = a;
uint8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t k = 0;
const __m256i vb_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.kernel_zero_point);
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepu8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb01), vb_zero_point);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 16));
const __m256i vxb23 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb23), vb_zero_point);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 32));
const __m256i vxb45 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb45), vb_zero_point);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 48));
const __m256i vxb67 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb67), vb_zero_point);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const uint8_t*) w + 64);
k += 8 * sizeof(uint8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packus_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epu8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (uint8_t*) ((uintptr_t) c0 + cn_stride);
a0 = (const uint8_t*) ((uintptr_t) a0 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (uint8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qu8_gemm_minmax_fp32_ukernel_3x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
const uint8_t* restrict a,
size_t a_stride,
const void* restrict w,
uint8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(kc % sizeof(uint8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
const uint8_t* a0 = a;
uint8_t* c0 = c;
const uint8_t* a1 = (const uint8_t*) ((uintptr_t) a0 + a_stride);
uint8_t* c1 = (uint8_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
a1 = a0;
c1 = c0;
}
const uint8_t* a2 = (const uint8_t*) ((uintptr_t) a1 + a_stride);
uint8_t* c2 = (uint8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
a2 = a1;
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t k = 0;
const __m256i vb_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.kernel_zero_point);
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepu8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepu8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepu8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb01), vb_zero_point);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 16));
const __m256i vxb23 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb23), vb_zero_point);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 32));
const __m256i vxb45 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb45), vb_zero_point);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 48));
const __m256i vxb67 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb67), vb_zero_point);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const uint8_t*) w + 64);
k += 8 * sizeof(uint8_t);
}
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packus_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epu8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
c0 = (uint8_t*) ((uintptr_t) c0 + cn_stride);
c1 = (uint8_t*) ((uintptr_t) c1 + cn_stride);
c2 = (uint8_t*) ((uintptr_t) c2 + cn_stride);
a0 = (const uint8_t*) ((uintptr_t) a0 - kc);
a1 = (const uint8_t*) ((uintptr_t) a1 - kc);
a2 = (const uint8_t*) ((uintptr_t) a2 - kc);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
_mm_storeu_si32(c1, vout_hi);
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
c0 += 4;
c1 += 4;
c2 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
c0 += 2;
c1 += 2;
c2 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (uint8_t) _mm_extract_epi8(vout_lo, 0);
*c1 = (uint8_t) _mm_extract_epi8(vout_hi, 0);
*c2 = (uint8_t) _mm_extract_epi8(vout_lo, 8);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qu8_igemm_minmax_fp32_ukernel_1x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const uint8_t** restrict a,
const void* restrict w,
uint8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const uint8_t* zero,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 1);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (1 * sizeof(void*)) == 0);
assert(a_offset % sizeof(uint8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
uint8_t* c0 = c;
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
w = (const int32_t*) w + 8;
size_t p = ks;
const __m256i vb_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.kernel_zero_point);
do {
const uint8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const uint8_t*) ((uintptr_t) a0 + a_offset);
}
a += 1;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepu8_epi16(va0);
a0 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb01), vb_zero_point);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 16));
const __m256i vxb23 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb23), vb_zero_point);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 32));
const __m256i vxb45 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb45), vb_zero_point);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 48));
const __m256i vxb67 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb67), vb_zero_point);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
w = (const void*) ((const uint8_t*) w + 64);
k += 8 * sizeof(uint8_t);
}
p -= 1 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc00x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc0x01234567), voutput_zero_point);
vacc00x01234567 = _mm256_permute4x64_epi64(vacc00x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packus_epi16(vacc00x01234567, vacc00x01234567);
vout = _mm256_max_epu8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storel_epi64((__m128i*) c0, vout_lo);
c0 = (uint8_t*) ((uintptr_t) c0 + cn_stride);
a = (const uint8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
_mm_storeu_si32(c0, vout_lo);
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c0 = (uint8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qu8_igemm_minmax_fp32_ukernel_3x8c8__avx2(
size_t mr,
size_t nc,
size_t kc,
size_t ks,
const uint8_t** restrict a,
const void* restrict w,
uint8_t* restrict c,
size_t cm_stride,
size_t cn_stride,
size_t a_offset,
const uint8_t* zero,
const union xnn_qu8_conv_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(mr != 0);
assert(mr <= 3);
assert(nc != 0);
assert(kc != 0);
assert(ks != 0);
assert(ks % (3 * sizeof(void*)) == 0);
assert(a_offset % sizeof(uint8_t) == 0);
assert(a != NULL);
assert(w != NULL);
assert(c != NULL);
kc = round_up_po2(kc, 8);
uint8_t* c0 = c;
uint8_t* c1 = (uint8_t*) ((uintptr_t) c0 + cm_stride);
if XNN_UNPREDICTABLE(mr < 2) {
c1 = c0;
}
uint8_t* c2 = (uint8_t*) ((uintptr_t) c1 + cm_stride);
if XNN_UNPREDICTABLE(mr <= 2) {
c2 = c1;
}
do {
const __m128i vbias0x0 = _mm_cvtsi32_si128(((const int*) w)[0]);
const __m128i vbias0x1 = _mm_cvtsi32_si128(((const int*) w)[1]);
__m256i vacc0x01 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x0), vbias0x1, 1);
const __m128i vbias0x2 = _mm_cvtsi32_si128(((const int*) w)[2]);
const __m128i vbias0x3 = _mm_cvtsi32_si128(((const int*) w)[3]);
__m256i vacc0x23 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x2), vbias0x3, 1);
const __m128i vbias0x4 = _mm_cvtsi32_si128(((const int*) w)[4]);
const __m128i vbias0x5 = _mm_cvtsi32_si128(((const int*) w)[5]);
__m256i vacc0x45 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x4), vbias0x5, 1);
const __m128i vbias0x6 = _mm_cvtsi32_si128(((const int*) w)[6]);
const __m128i vbias0x7 = _mm_cvtsi32_si128(((const int*) w)[7]);
__m256i vacc0x67 = _mm256_inserti128_si256(_mm256_castsi128_si256(vbias0x6), vbias0x7, 1);
__m256i vacc1x01 = vacc0x01;
__m256i vacc1x23 = vacc0x23;
__m256i vacc1x45 = vacc0x45;
__m256i vacc1x67 = vacc0x67;
__m256i vacc2x01 = vacc0x01;
__m256i vacc2x23 = vacc0x23;
__m256i vacc2x45 = vacc0x45;
__m256i vacc2x67 = vacc0x67;
w = (const int32_t*) w + 8;
size_t p = ks;
const __m256i vb_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.kernel_zero_point);
do {
const uint8_t* restrict a0 = a[0];
if XNN_UNPREDICTABLE(a0 != zero) {
a0 = (const uint8_t*) ((uintptr_t) a0 + a_offset);
}
const uint8_t* restrict a1 = a[1];
if XNN_UNPREDICTABLE(a1 != zero) {
a1 = (const uint8_t*) ((uintptr_t) a1 + a_offset);
}
const uint8_t* restrict a2 = a[2];
if XNN_UNPREDICTABLE(a2 != zero) {
a2 = (const uint8_t*) ((uintptr_t) a2 + a_offset);
}
a += 3;
size_t k = 0;
while (k < kc) {
const __m128i va0 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a0));
const __m256i vxa0 = _mm256_cvtepu8_epi16(va0);
a0 += 8;
const __m128i va1 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a1));
const __m256i vxa1 = _mm256_cvtepu8_epi16(va1);
a1 += 8;
const __m128i va2 = _mm_broadcastq_epi64(_mm_loadl_epi64((const __m128i*) a2));
const __m256i vxa2 = _mm256_cvtepu8_epi16(va2);
a2 += 8;
const __m128i vb01 = _mm_load_si128((const __m128i*) w);
const __m256i vxb01 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb01), vb_zero_point);
vacc0x01 = _mm256_add_epi32(vacc0x01, _mm256_madd_epi16(vxa0, vxb01));
vacc1x01 = _mm256_add_epi32(vacc1x01, _mm256_madd_epi16(vxa1, vxb01));
vacc2x01 = _mm256_add_epi32(vacc2x01, _mm256_madd_epi16(vxa2, vxb01));
const __m128i vb23 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 16));
const __m256i vxb23 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb23), vb_zero_point);
vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23));
vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23));
vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23));
const __m128i vb45 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 32));
const __m256i vxb45 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb45), vb_zero_point);
vacc0x45 = _mm256_add_epi32(vacc0x45, _mm256_madd_epi16(vxa0, vxb45));
vacc1x45 = _mm256_add_epi32(vacc1x45, _mm256_madd_epi16(vxa1, vxb45));
vacc2x45 = _mm256_add_epi32(vacc2x45, _mm256_madd_epi16(vxa2, vxb45));
const __m128i vb67 = _mm_load_si128((const __m128i*) ((const uint8_t*) w + 48));
const __m256i vxb67 = _mm256_sub_epi16(_mm256_cvtepu8_epi16(vb67), vb_zero_point);
vacc0x67 = _mm256_add_epi32(vacc0x67, _mm256_madd_epi16(vxa0, vxb67));
vacc1x67 = _mm256_add_epi32(vacc1x67, _mm256_madd_epi16(vxa1, vxb67));
vacc2x67 = _mm256_add_epi32(vacc2x67, _mm256_madd_epi16(vxa2, vxb67));
w = (const void*) ((const uint8_t*) w + 64);
k += 8 * sizeof(uint8_t);
}
p -= 3 * sizeof(void*);
} while (p != 0);
const __m256i vacc0x0213 = _mm256_hadd_epi32(vacc0x01, vacc0x23);
const __m256i vacc0x4657 = _mm256_hadd_epi32(vacc0x45, vacc0x67);
const __m256i vacc1x0213 = _mm256_hadd_epi32(vacc1x01, vacc1x23);
const __m256i vacc1x4657 = _mm256_hadd_epi32(vacc1x45, vacc1x67);
const __m256i vacc2x0213 = _mm256_hadd_epi32(vacc2x01, vacc2x23);
const __m256i vacc2x4657 = _mm256_hadd_epi32(vacc2x45, vacc2x67);
const __m256i vacc0x02461357 = _mm256_hadd_epi32(vacc0x0213, vacc0x4657);
const __m256i vacc1x02461357 = _mm256_hadd_epi32(vacc1x0213, vacc1x4657);
const __m256i vacc2x02461357 = _mm256_hadd_epi32(vacc2x0213, vacc2x4657);
const __m256i vpermute_mask = _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0);
__m256i vacc0x01234567 = _mm256_permutevar8x32_epi32(vacc0x02461357, vpermute_mask);
__m256i vacc1x01234567 = _mm256_permutevar8x32_epi32(vacc1x02461357, vpermute_mask);
__m256i vacc2x01234567 = _mm256_permutevar8x32_epi32(vacc2x02461357, vpermute_mask);
__m256 vscaled0x01234567 = _mm256_cvtepi32_ps(vacc0x01234567);
__m256 vscaled1x01234567 = _mm256_cvtepi32_ps(vacc1x01234567);
__m256 vscaled2x01234567 = _mm256_cvtepi32_ps(vacc2x01234567);
const __m256 vscale = _mm256_load_ps(params->fp32_avx2.scale);
vscaled0x01234567 = _mm256_mul_ps(vscaled0x01234567, vscale);
vscaled1x01234567 = _mm256_mul_ps(vscaled1x01234567, vscale);
vscaled2x01234567 = _mm256_mul_ps(vscaled2x01234567, vscale);
const __m256 voutput_max_less_zero_point = _mm256_load_ps(params->fp32_avx2.output_max_less_zero_point);
vscaled0x01234567 = _mm256_min_ps(vscaled0x01234567, voutput_max_less_zero_point);
vscaled1x01234567 = _mm256_min_ps(vscaled1x01234567, voutput_max_less_zero_point);
vscaled2x01234567 = _mm256_min_ps(vscaled2x01234567, voutput_max_less_zero_point);
vacc0x01234567 = _mm256_cvtps_epi32(vscaled0x01234567);
vacc1x01234567 = _mm256_cvtps_epi32(vscaled1x01234567);
vacc2x01234567 = _mm256_cvtps_epi32(vscaled2x01234567);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->fp32_avx2.output_zero_point);
__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), voutput_zero_point);
__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), voutput_zero_point);
vacc01x01234567 = _mm256_permute4x64_epi64(vacc01x01234567, _MM_SHUFFLE(3, 1, 2, 0));
vacc22x01234567 = _mm256_permute4x64_epi64(vacc22x01234567, _MM_SHUFFLE(3, 1, 2, 0));
__m256i vout = _mm256_packus_epi16(vacc01x01234567, vacc22x01234567);
vout = _mm256_max_epu8(vout, _mm256_load_si256((const __m256i*) params->fp32_avx2.output_min));
__m128i vout_lo = _mm256_castsi256_si128(vout);
__m128i vout_hi = _mm256_extracti128_si256(vout, 1);
if (nc >= 8) {
_mm_storeh_pi((__m64*) c2, _mm_castsi128_ps(vout_lo));
_mm_storel_epi64((__m128i*) c1, vout_hi);
_mm_storel_epi64((__m128i*) c0, vout_lo);
c2 = (uint8_t*) ((uintptr_t) c2 + cn_stride);
c1 = (uint8_t*) ((uintptr_t) c1 + cn_stride);
c0 = (uint8_t*) ((uintptr_t) c0 + cn_stride);
a = (const uint8_t**restrict) ((uintptr_t) a - ks);
nc -= 8;
} else {
if (nc & 4) {
unaligned_store_u32(c2, (uint32_t) _mm_extract_epi32(vout_lo, 2));
_mm_storeu_si32(c1, vout_hi);
_mm_storeu_si32(c0, vout_lo);
c2 += 4;
c1 += 4;
c0 += 4;
vout_lo = _mm_srli_epi64(vout_lo, 32);
vout_hi = _mm_srli_epi64(vout_hi, 32);
}
if (nc & 2) {
unaligned_store_u16(c2, (uint16_t) _mm_extract_epi16(vout_lo, 4));
unaligned_store_u16(c1, (uint16_t) _mm_extract_epi16(vout_hi, 0));
unaligned_store_u16(c0, (uint16_t) _mm_extract_epi16(vout_lo, 0));
c2 += 2;
c1 += 2;
c0 += 2;
vout_lo = _mm_srli_epi32(vout_lo, 16);
vout_hi = _mm_srli_epi32(vout_hi, 16);
}
if (nc & 1) {
*c2 = (uint8_t) _mm_extract_epi8(vout_lo, 8);
*c1 = (uint8_t) _mm_extract_epi8(vout_hi, 0);
*c0 = (uint8_t) _mm_extract_epi8(vout_lo, 0);
}
nc = 0;
}
} while (nc != 0);
}
void xnn_qu8_vadd_minmax_ukernel__avx2_mul32_ld64_x16(
size_t n,
const uint8_t* input_a,
const uint8_t* input_b,
uint8_t* output,
const union xnn_qu8_add_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
const __m256i vbias = _mm256_load_si256((const __m256i*) params->avx2.bias);
const __m256i va_multiplier = _mm256_load_si256((const __m256i*) params->avx2.a_multiplier);
const __m256i vb_multiplier = _mm256_load_si256((const __m256i*) params->avx2.b_multiplier);
const __m128i vshift = _mm_load_si128((const __m128i*) params->avx2.shift);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->avx2.output_min);
const __m128i voutput_max = _mm_load_si128((const __m128i*) params->avx2.output_max);
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
const __m256i va01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i vb01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_b));
const __m256i va89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (input_a + 8)));
const __m256i vb89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (input_b + 8)));
input_a += 16;
input_b += 16;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
__m256i vacc89ABCDEF = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va89ABCDEF, va_multiplier));
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vb01234567, vb_multiplier));
vacc89ABCDEF = _mm256_add_epi32(vacc89ABCDEF, _mm256_mullo_epi32(vb89ABCDEF, vb_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
vacc89ABCDEF = _mm256_sra_epi32(vacc89ABCDEF, vshift);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packus_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
vout0123456789ABCDEF = _mm_max_epu8(vout0123456789ABCDEF, voutput_min);
vout0123456789ABCDEF = _mm_min_epu8(vout0123456789ABCDEF, voutput_max);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(n != 0) {
do {
const __m256i va01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i vb01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_b));
input_a += 8;
input_b += 8;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
vacc01234567 = _mm256_add_epi32(vacc01234567, _mm256_mullo_epi32(vb01234567, vb_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), _mm256_castsi256_si128(voutput_zero_point));
__m128i vout0123456701234567 = _mm_packus_epi16(vout01234567, vout01234567);
vout0123456701234567 = _mm_max_epu8(vout0123456701234567, voutput_min);
vout0123456701234567 = _mm_min_epu8(vout0123456701234567, voutput_max);
if XNN_LIKELY(n >= (8 * sizeof(uint8_t))) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
n -= 8 * sizeof(uint8_t);
} else {
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_si32(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storeu_si16(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (n & (1 * sizeof(uint8_t))) {
*output = (uint8_t) _mm_extract_epi8(vout0123456701234567, 0);
}
n = 0;
}
} while (n != 0);
}
}
void xnn_qu8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16(
size_t n,
const uint8_t* input_a,
const uint8_t* input_b,
uint8_t* output,
const union xnn_qu8_add_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
const __m256i va_multiplier = _mm256_load_si256((const __m256i*) params->avx2.a_multiplier);
const __m128i vshift = _mm_load_si128((const __m128i*) params->avx2.shift);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
const __m128i voutput_min = _mm_load_si128((const __m128i*) params->avx2.output_min);
const __m128i voutput_max = _mm_load_si128((const __m128i*) params->avx2.output_max);
const __m256i vbias = _mm256_add_epi32(
_mm256_broadcastd_epi32(_mm_cvtsi32_si128(params->avx2.b_multiplier[0] * (int32_t) *input_b)),
_mm256_load_si256((const __m256i*) params->avx2.bias));
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
const __m256i va01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
const __m256i va89ABCDEF = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) (input_a + 8)));
input_a += 16;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
__m256i vacc89ABCDEF = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va89ABCDEF, va_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
vacc89ABCDEF = _mm256_sra_epi32(vacc89ABCDEF, vshift);
__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), voutput_zero_point);
__m128i vout0123456789ABCDEF = _mm_shuffle_epi32(_mm_packus_epi16(_mm256_castsi256_si128(vout012389AB4567CDEF), _mm256_extracti128_si256(vout012389AB4567CDEF, 1)), _MM_SHUFFLE(3, 1, 2, 0));
vout0123456789ABCDEF = _mm_max_epu8(vout0123456789ABCDEF, voutput_min);
vout0123456789ABCDEF = _mm_min_epu8(vout0123456789ABCDEF, voutput_max);
_mm_storeu_si128((__m128i*) output, vout0123456789ABCDEF);
output += 16;
}
if XNN_UNLIKELY(n != 0) {
do {
const __m256i va01234567 = _mm256_cvtepu8_epi32(_mm_loadl_epi64((const __m128i*) input_a));
input_a += 8;
__m256i vacc01234567 = _mm256_add_epi32(vbias, _mm256_mullo_epi32(va01234567, va_multiplier));
vacc01234567 = _mm256_sra_epi32(vacc01234567, vshift);
__m128i vout01234567 = _mm_adds_epi16(_mm_packs_epi32(_mm256_castsi256_si128(vacc01234567), _mm256_extracti128_si256(vacc01234567, 1)), _mm256_castsi256_si128(voutput_zero_point));
__m128i vout0123456701234567 = _mm_packus_epi16(vout01234567, vout01234567);
vout0123456701234567 = _mm_max_epu8(vout0123456701234567, voutput_min);
vout0123456701234567 = _mm_min_epu8(vout0123456701234567, voutput_max);
if XNN_LIKELY(n >= (8 * sizeof(uint8_t))) {
_mm_storel_epi64((__m128i*) output, vout0123456701234567);
output += 8;
n -= 8 * sizeof(uint8_t);
} else {
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_si32(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi64(vout0123456701234567, 32);
output += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storeu_si16(output, vout0123456701234567);
vout0123456701234567 = _mm_srli_epi32(vout0123456701234567, 16);
output += 2;
}
if (n & (1 * sizeof(uint8_t))) {
*output = (uint8_t) _mm_extract_epi8(vout0123456701234567, 0);
}
n = 0;
}
} while (n != 0);
}
}
void xnn_qu8_vcvt_ukernel__avx2_x32(
size_t n,
const uint8_t* x,
uint8_t* y,
const union xnn_qu8_cvt_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(uint8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vinput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.input_zero_point);
const __m256i vmultiplier = _mm256_load_si256((const __m256i*) params->avx2.multiplier);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
for (; n >= 32 * sizeof(uint8_t); n -= 32 * sizeof(uint8_t)) {
__m256i vacc0 = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vacc1 = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) (x + 16)));
x += 32;
vacc0 = _mm256_sub_epi16(vinput_zero_point, vacc0);
vacc1 = _mm256_sub_epi16(vinput_zero_point, vacc1);
vacc0 = _mm256_slli_epi16(vacc0, 7);
vacc1 = _mm256_slli_epi16(vacc1, 7);
vacc0 = _mm256_mulhrs_epi16(vacc0, vmultiplier);
vacc1 = _mm256_mulhrs_epi16(vacc1, vmultiplier);
vacc0 = _mm256_adds_epi16(vacc0, voutput_zero_point);
vacc1 = _mm256_adds_epi16(vacc1, voutput_zero_point);
__m256i vy0 = _mm256_packus_epi16(vacc0, vacc1);
vy0 = _mm256_permute4x64_epi64(vy0, _MM_SHUFFLE(3, 1, 2, 0));
_mm256_storeu_si256((__m256i*) y, vy0);
y += 32;
}
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
__m256i vacc = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
x += 16;
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
const __m128i vy = _mm_packus_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
_mm_storeu_si128((__m128i*) y, vy);
y += 16;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(uint8_t));
assert(n <= 15 * sizeof(uint8_t));
__m256i vacc = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
__m128i vy = _mm_packus_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
if (n & (8 * sizeof(uint8_t))) {
_mm_storel_epi64((__m128i*) y, vy);
vy = _mm_unpackhi_epi64(vy, vy);
y += 8;
}
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_si32(y, vy);
vy = _mm_srli_epi64(vy, 32);
y += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storeu_si16(y, vy);
vy = _mm_srli_epi32(vy, 16);
y += 2;
}
if (n & (1 * sizeof(uint8_t))) {
*y = (uint8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_qu8_vlrelu_ukernel__avx2_x32(
size_t n,
const uint8_t* x,
uint8_t* y,
const union xnn_qu8_lrelu_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
assert(n != 0);
assert(n % sizeof(uint8_t) == 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vinput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.input_zero_point);
const __m256i vpositive_multiplier = _mm256_load_si256((const __m256i*) params->avx2.positive_multiplier);
const __m256i vnegative_multiplier = _mm256_load_si256((const __m256i*) params->avx2.negative_multiplier);
const __m256i voutput_zero_point = _mm256_load_si256((const __m256i*) params->avx2.output_zero_point);
for (; n >= 32 * sizeof(uint8_t); n -= 32 * sizeof(uint8_t)) {
__m256i vacc0 = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vacc1 = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) (x + 16)));
x += 32;
__m256i vmultiplier0 = _mm256_cmpgt_epi16(vacc0, vinput_zero_point);
vacc0 = _mm256_sub_epi16(vinput_zero_point, vacc0);
__m256i vmultiplier1 = _mm256_cmpgt_epi16(vacc1, vinput_zero_point);
vacc1 = _mm256_sub_epi16(vinput_zero_point, vacc1);
vmultiplier0 = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier0);
vacc0 = _mm256_slli_epi16(vacc0, 7);
vmultiplier1 = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier1);
vacc1 = _mm256_slli_epi16(vacc1, 7);
vacc0 = _mm256_mulhrs_epi16(vacc0, vmultiplier0);
vacc1 = _mm256_mulhrs_epi16(vacc1, vmultiplier1);
vacc0 = _mm256_adds_epi16(vacc0, voutput_zero_point);
vacc1 = _mm256_adds_epi16(vacc1, voutput_zero_point);
__m256i vy0 = _mm256_packus_epi16(vacc0, vacc1);
vy0 = _mm256_permute4x64_epi64(vy0, _MM_SHUFFLE(3, 1, 2, 0));
_mm256_storeu_si256((__m256i*) y, vy0);
y += 32;
}
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
__m256i vacc = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vmultiplier = _mm256_cmpgt_epi16(vacc, vinput_zero_point);
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vmultiplier = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
x += 16;
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
const __m128i vy = _mm_packus_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
_mm_storeu_si128((__m128i*) y, vy);
y += 16;
}
if XNN_UNLIKELY(n != 0) {
assert(n >= 1 * sizeof(uint8_t));
assert(n <= 15 * sizeof(uint8_t));
__m256i vacc = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i*) x));
__m256i vmultiplier = _mm256_cmpgt_epi16(vacc, vinput_zero_point);
vacc = _mm256_sub_epi16(vinput_zero_point, vacc);
vmultiplier = _mm256_blendv_epi8(vnegative_multiplier, vpositive_multiplier, vmultiplier);
vacc = _mm256_slli_epi16(vacc, 7);
vacc = _mm256_mulhrs_epi16(vacc, vmultiplier);
vacc = _mm256_adds_epi16(vacc, voutput_zero_point);
const __m128i vacc_hi = _mm256_extracti128_si256(vacc, 1);
__m128i vy = _mm_packus_epi16(_mm256_castsi256_si128(vacc), vacc_hi);
if (n & (8 * sizeof(uint8_t))) {
_mm_storel_epi64((__m128i*) y, vy);
vy = _mm_unpackhi_epi64(vy, vy);
y += 8;
}
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_si32(y, vy);
vy = _mm_srli_epi64(vy, 32);
y += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storeu_si16(y, vy);
vy = _mm_srli_epi32(vy, 16);
y += 2;
}
if (n & (1 * sizeof(uint8_t))) {
*y = (uint8_t) _mm_extract_epi8(vy, 0);
}
}
}
void xnn_x8_lut_ukernel__avx2_x128(
size_t n,
const uint8_t* x,
uint8_t* y,
const uint8_t t[restrict XNN_MIN_ELEMENTS(256)])
{
assert(n != 0);
assert(x != NULL);
assert(y != NULL);
const __m256i vt0 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) t));
const __m256i vt1 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 16)));
const __m256i vt2 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 32)));
const __m256i vt3 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 48)));
const __m256i vt4 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 64)));
const __m256i vt5 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 80)));
const __m256i vt6 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 96)));
const __m256i vt7 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 112)));
const __m256i vt8 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 128)));
const __m256i vt9 = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 144)));
const __m256i vtA = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 160)));
const __m256i vtB = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 176)));
const __m256i vtC = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 192)));
const __m256i vtD = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 208)));
const __m256i vtE = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 224)));
const __m256i vtF = _mm256_broadcastsi128_si256(_mm_load_si128((const __m128i*) (t + 240)));
const __m256i vtable0 = vt0;
const __m256i vtable1 = _mm256_xor_si256(vt0, vt1);
const __m256i vtable2 = _mm256_xor_si256(vt1, vt2);
const __m256i vtable3 = _mm256_xor_si256(vt2, vt3);
const __m256i vtable4 = _mm256_xor_si256(vt3, vt4);
const __m256i vtable5 = _mm256_xor_si256(vt4, vt5);
const __m256i vtable6 = _mm256_xor_si256(vt5, vt6);
const __m256i vtable7 = _mm256_xor_si256(vt6, vt7);
const __m256i vtable8 = _mm256_xor_si256(_mm256_xor_si256(vt7, vt8), vtable0);
const __m256i vtable9 = _mm256_xor_si256(_mm256_xor_si256(vt8, vt9), vtable1);
const __m256i vtableA = _mm256_xor_si256(_mm256_xor_si256(vt9, vtA), vtable2);
const __m256i vtableB = _mm256_xor_si256(_mm256_xor_si256(vtA, vtB), vtable3);
const __m256i vtableC = _mm256_xor_si256(_mm256_xor_si256(vtB, vtC), vtable4);
const __m256i vtableD = _mm256_xor_si256(_mm256_xor_si256(vtC, vtD), vtable5);
const __m256i vtableE = _mm256_xor_si256(_mm256_xor_si256(vtD, vtE), vtable6);
const __m256i vtableF = _mm256_xor_si256(_mm256_xor_si256(vtE, vtF), vtable7);
const __m256i voffset = _mm256_set1_epi8(16);
for (; n >= 128 * sizeof(uint8_t); n -= 128 * sizeof(uint8_t)) {
__m256i vx0 = _mm256_loadu_si256((const __m256i*) x);
__m256i vx1 = _mm256_loadu_si256((const __m256i*) (x + 32));
__m256i vx2 = _mm256_loadu_si256((const __m256i*) (x + 64));
__m256i vx3 = _mm256_loadu_si256((const __m256i*) (x + 96));
x += 128;
__m256i vy0 = _mm256_shuffle_epi8(vtable0, vx0);
__m256i vy1 = _mm256_shuffle_epi8(vtable0, vx1);
__m256i vy2 = _mm256_shuffle_epi8(vtable0, vx2);
__m256i vy3 = _mm256_shuffle_epi8(vtable0, vx3);
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable1, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable1, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable1, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable1, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable2, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable2, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable2, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable2, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable3, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable3, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable3, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable3, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable4, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable4, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable4, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable4, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable5, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable5, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable5, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable5, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable6, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable6, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable6, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable6, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable7, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable7, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable7, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable7, vx3));
vx0 = _mm256_sub_epi8(vx0, voffset);
vx1 = _mm256_sub_epi8(vx1, voffset);
vx2 = _mm256_sub_epi8(vx2, voffset);
vx3 = _mm256_sub_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable8, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable8, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable8, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable8, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtable9, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtable9, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtable9, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtable9, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableA, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableA, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableA, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableA, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableB, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableB, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableB, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableB, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableC, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableC, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableC, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableC, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableD, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableD, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableD, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableD, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableE, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableE, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableE, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableE, vx3));
vx0 = _mm256_subs_epi8(vx0, voffset);
vx1 = _mm256_subs_epi8(vx1, voffset);
vx2 = _mm256_subs_epi8(vx2, voffset);
vx3 = _mm256_subs_epi8(vx3, voffset);
vy0 = _mm256_xor_si256(vy0, _mm256_shuffle_epi8(vtableF, vx0));
vy1 = _mm256_xor_si256(vy1, _mm256_shuffle_epi8(vtableF, vx1));
vy2 = _mm256_xor_si256(vy2, _mm256_shuffle_epi8(vtableF, vx2));
vy3 = _mm256_xor_si256(vy3, _mm256_shuffle_epi8(vtableF, vx3));
_mm256_storeu_si256((__m256i*) y, vy0);
_mm256_storeu_si256((__m256i*) (y + 32), vy1);
_mm256_storeu_si256((__m256i*) (y + 64), vy2);
_mm256_storeu_si256((__m256i*) (y + 96), vy3);
y += 128;
}
for (; n >= 16 * sizeof(uint8_t); n -= 16 * sizeof(uint8_t)) {
__m128i vx = _mm_loadu_si128((const __m128i*) x);
x += 16;
__m128i vy = _mm_shuffle_epi8(_mm256_castsi256_si128(vtable0), vx);
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable1), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable2), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable3), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable4), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable5), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable6), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable7), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable8), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable9), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableA), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableB), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableC), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableD), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableE), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableF), vx));
_mm_storeu_si128((__m128i*) y, vy);
y += 16;
}
if XNN_UNLIKELY(n != 0) {
__m128i vx = _mm_loadu_si128((const __m128i*) x);
__m128i vy = _mm_shuffle_epi8(_mm256_castsi256_si128(vtable0), vx);
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable1), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable2), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable3), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable4), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable5), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable6), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable7), vx));
vx = _mm_sub_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable8), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtable9), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableA), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableB), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableC), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableD), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableE), vx));
vx = _mm_subs_epi8(vx, _mm256_castsi256_si128(voffset));
vy = _mm_xor_si128(vy, _mm_shuffle_epi8(_mm256_castsi256_si128(vtableF), vx));
if (n & (8 * sizeof(uint8_t))) {
_mm_storel_epi64((__m128i*) y, vy);
vy = _mm_unpackhi_epi64(vy, vy);
y += 8;
}
if (n & (4 * sizeof(uint8_t))) {
_mm_storeu_si32(y, vy);
vy = _mm_srli_epi64(vy, 32);
y += 4;
}
if (n & (2 * sizeof(uint8_t))) {
_mm_storeu_si16(y, vy);
vy = _mm_srli_epi32(vy, 16);
y += 2;
}
if (n & (1 * sizeof(uint8_t))) {
*y = (uint8_t) _mm_extract_epi8(vy, 0);
}
}
}