blob: 64cd810f77213849895ddc6c4476032117df23f1 [file] [log] [blame]
/*
* Copyright (c) 2017, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
#ifndef THIRD_PARTY_SVT_AV1_CONVOLVE_2D_AVX2_H_
#define THIRD_PARTY_SVT_AV1_CONVOLVE_2D_AVX2_H_
#include "convolve_avx2.h"
static void convolve_2d_sr_hor_2tap_avx2(
const uint8_t *const src, const int32_t src_stride, const int32_t w,
const int32_t h, const InterpFilterParams *const filter_params_x,
const int32_t subpel_x_q4, int16_t *const im_block) {
const uint8_t *src_ptr = src;
int32_t y = h;
int16_t *im = im_block;
if (w <= 8) {
__m128i coeffs_128;
prepare_half_coeffs_2tap_ssse3(filter_params_x, subpel_x_q4, &coeffs_128);
if (w == 2) {
do {
const __m128i r =
x_convolve_2tap_2x2_sse4_1(src_ptr, src_stride, &coeffs_128);
xy_x_round_store_2x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 2;
y -= 2;
} while (y);
} else if (w == 4) {
do {
const __m128i r =
x_convolve_2tap_4x2_ssse3(src_ptr, src_stride, &coeffs_128);
xy_x_round_store_4x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 4;
y -= 2;
} while (y);
} else {
assert(w == 8);
do {
__m128i r[2];
x_convolve_2tap_8x2_ssse3(src_ptr, src_stride, &coeffs_128, r);
xy_x_round_store_8x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 8;
y -= 2;
} while (y);
}
} else {
__m256i coeffs_256;
prepare_half_coeffs_2tap_avx2(filter_params_x, subpel_x_q4, &coeffs_256);
if (w == 16) {
do {
__m256i r[2];
x_convolve_2tap_16x2_avx2(src_ptr, src_stride, &coeffs_256, r);
xy_x_round_store_32_avx2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 16;
y -= 2;
} while (y);
} else if (w == 32) {
do {
xy_x_2tap_32_avx2(src_ptr, &coeffs_256, im);
src_ptr += src_stride;
im += 32;
} while (--y);
} else if (w == 64) {
do {
xy_x_2tap_32_avx2(src_ptr + 0 * 32, &coeffs_256, im + 0 * 32);
xy_x_2tap_32_avx2(src_ptr + 1 * 32, &coeffs_256, im + 1 * 32);
src_ptr += src_stride;
im += 64;
} while (--y);
} else {
assert(w == 128);
do {
xy_x_2tap_32_avx2(src_ptr + 0 * 32, &coeffs_256, im + 0 * 32);
xy_x_2tap_32_avx2(src_ptr + 1 * 32, &coeffs_256, im + 1 * 32);
xy_x_2tap_32_avx2(src_ptr + 2 * 32, &coeffs_256, im + 2 * 32);
xy_x_2tap_32_avx2(src_ptr + 3 * 32, &coeffs_256, im + 3 * 32);
src_ptr += src_stride;
im += 128;
} while (--y);
}
}
}
static void convolve_2d_sr_hor_4tap_ssse3(
const uint8_t *const src, const int32_t src_stride, const int32_t w,
const int32_t h, const InterpFilterParams *const filter_params_x,
const int32_t subpel_x_q4, int16_t *const im_block) {
const uint8_t *src_ptr = src - 1;
int32_t y = h;
int16_t *im = im_block;
if (w <= 4) {
__m128i coeffs_128[2];
prepare_half_coeffs_4tap_ssse3(filter_params_x, subpel_x_q4, coeffs_128);
if (w == 2) {
do {
const __m128i r =
x_convolve_4tap_2x2_ssse3(src_ptr, src_stride, coeffs_128);
xy_x_round_store_2x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 2;
y -= 2;
} while (y);
} else if (w == 4) {
do {
const __m128i r =
x_convolve_4tap_4x2_ssse3(src_ptr, src_stride, coeffs_128);
xy_x_round_store_4x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 4;
y -= 2;
} while (y);
}
} else {
// TODO(chiyotsai@google.com): Add better optimization
__m256i coeffs_256[2], filt_256[2];
prepare_half_coeffs_4tap_avx2(filter_params_x, subpel_x_q4, coeffs_256);
filt_256[0] = _mm256_load_si256((__m256i const *)filt1_global_avx2);
filt_256[1] = _mm256_load_si256((__m256i const *)filt2_global_avx2);
if (w == 8) {
do {
__m256i res =
x_convolve_4tap_8x2_avx2(src_ptr, src_stride, coeffs_256, filt_256);
xy_x_round_store_8x2_avx2(res, im);
src_ptr += 2 * src_stride;
im += 2 * 8;
y -= 2;
} while (y);
} else if (w == 16) {
do {
__m256i r[2];
x_convolve_4tap_16x2_avx2(src_ptr, src_stride, coeffs_256, filt_256, r);
xy_x_round_store_32_avx2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 16;
y -= 2;
} while (y);
} else if (w == 32) {
do {
xy_x_4tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
src_ptr += src_stride;
im += 32;
} while (--y);
} else if (w == 64) {
do {
xy_x_4tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_4tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
src_ptr += src_stride;
im += 64;
} while (--y);
} else {
assert(w == 128);
do {
xy_x_4tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_4tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
xy_x_4tap_32_avx2(src_ptr + 64, coeffs_256, filt_256, im + 64);
xy_x_4tap_32_avx2(src_ptr + 96, coeffs_256, filt_256, im + 96);
src_ptr += src_stride;
im += 128;
} while (--y);
}
}
}
static void convolve_2d_sr_hor_6tap_avx2(
const uint8_t *const src, const int32_t src_stride, const int32_t w,
const int32_t h, const InterpFilterParams *const filter_params_x,
const int32_t subpel_x_q4, int16_t *const im_block) {
const uint8_t *src_ptr = src - 2;
int32_t y = h;
int16_t *im = im_block;
if (w <= 4) {
__m128i coeffs_128[3];
prepare_half_coeffs_6tap_ssse3(filter_params_x, subpel_x_q4, coeffs_128);
if (w == 2) {
do {
const __m128i r =
x_convolve_6tap_2x2_ssse3(src_ptr, src_stride, coeffs_128);
xy_x_round_store_2x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 2;
y -= 2;
} while (y);
} else if (w == 4) {
do {
const __m128i r =
x_convolve_6tap_4x2_ssse3(src_ptr, src_stride, coeffs_128);
xy_x_round_store_4x2_sse2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 4;
y -= 2;
} while (y);
}
} else {
__m256i coeffs_256[3], filt_256[3];
filt_256[0] = _mm256_loadu_si256((__m256i const *)filt1_global_avx2);
filt_256[1] = _mm256_loadu_si256((__m256i const *)filt2_global_avx2);
filt_256[2] = _mm256_loadu_si256((__m256i const *)filt3_global_avx2);
prepare_half_coeffs_6tap_avx2(filter_params_x, subpel_x_q4, coeffs_256);
if (w == 8) {
do {
const __m256i res =
x_convolve_6tap_8x2_avx2(src_ptr, src_stride, coeffs_256, filt_256);
xy_x_round_store_8x2_avx2(res, im);
src_ptr += 2 * src_stride;
im += 2 * 8;
y -= 2;
} while (y);
} else if (w == 16) {
do {
__m256i r[2];
x_convolve_6tap_16x2_avx2(src_ptr, src_stride, coeffs_256, filt_256, r);
xy_x_round_store_32_avx2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 16;
y -= 2;
} while (y);
} else if (w == 32) {
do {
xy_x_6tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
src_ptr += src_stride;
im += 32;
} while (--y);
} else if (w == 64) {
do {
xy_x_6tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_6tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
src_ptr += src_stride;
im += 64;
} while (--y);
} else {
assert(w == 128);
do {
xy_x_6tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_6tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
xy_x_6tap_32_avx2(src_ptr + 64, coeffs_256, filt_256, im + 64);
xy_x_6tap_32_avx2(src_ptr + 96, coeffs_256, filt_256, im + 96);
src_ptr += src_stride;
im += 128;
} while (--y);
}
}
}
static void convolve_2d_sr_hor_8tap_avx2(
const uint8_t *const src, const int32_t src_stride, const int32_t w,
const int32_t h, const InterpFilterParams *const filter_params_x,
const int32_t subpel_x_q4, int16_t *const im_block) {
const uint8_t *src_ptr = src - 3;
int32_t y = h;
int16_t *im = im_block;
__m256i coeffs_256[4], filt_256[4];
filt_256[0] = _mm256_loadu_si256((__m256i const *)filt1_global_avx2);
filt_256[1] = _mm256_loadu_si256((__m256i const *)filt2_global_avx2);
filt_256[2] = _mm256_loadu_si256((__m256i const *)filt3_global_avx2);
filt_256[3] = _mm256_loadu_si256((__m256i const *)filt4_global_avx2);
prepare_half_coeffs_8tap_avx2(filter_params_x, subpel_x_q4, coeffs_256);
if (w == 8) {
do {
const __m256i res =
x_convolve_8tap_8x2_avx2(src_ptr, src_stride, coeffs_256, filt_256);
xy_x_round_store_8x2_avx2(res, im);
src_ptr += 2 * src_stride;
im += 2 * 8;
y -= 2;
} while (y);
} else if (w == 16) {
do {
__m256i r[2];
x_convolve_8tap_16x2_avx2(src_ptr, src_stride, coeffs_256, filt_256, r);
xy_x_round_store_32_avx2(r, im);
src_ptr += 2 * src_stride;
im += 2 * 16;
y -= 2;
} while (y);
} else if (w == 32) {
do {
xy_x_8tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
src_ptr += src_stride;
im += 32;
} while (--y);
} else if (w == 64) {
do {
xy_x_8tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_8tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
src_ptr += src_stride;
im += 64;
} while (--y);
} else {
assert(w == 128);
do {
xy_x_8tap_32_avx2(src_ptr, coeffs_256, filt_256, im);
xy_x_8tap_32_avx2(src_ptr + 32, coeffs_256, filt_256, im + 32);
xy_x_8tap_32_avx2(src_ptr + 64, coeffs_256, filt_256, im + 64);
xy_x_8tap_32_avx2(src_ptr + 96, coeffs_256, filt_256, im + 96);
src_ptr += src_stride;
im += 128;
} while (--y);
}
}
static void convolve_2d_sr_ver_2tap_avx2(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride) {
const int16_t *im = im_block;
int32_t y = h;
if (w <= 4) {
__m128i coeffs_128;
prepare_coeffs_2tap_sse2(filter_params_y, subpel_y_q4, &coeffs_128);
if (w == 2) {
__m128i s_32[2];
s_32[0] = _mm_cvtsi32_si128(*(int32_t *)im);
do {
const __m128i res = xy_y_convolve_2tap_2x2_sse2(im, s_32, &coeffs_128);
xy_y_round_store_2x2_sse2(res, dst, dst_stride);
im += 2 * 2;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m128i s_64[2], r[2];
assert(w == 4);
s_64[0] = _mm_loadl_epi64((__m128i *)im);
do {
xy_y_convolve_2tap_4x2_sse2(im, s_64, &coeffs_128, r);
r[0] = xy_y_round_sse2(r[0]);
r[1] = xy_y_round_sse2(r[1]);
const __m128i rr = _mm_packs_epi32(r[0], r[1]);
pack_store_4x2_sse2(rr, dst, dst_stride);
im += 2 * 4;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else {
__m256i coeffs_256;
prepare_coeffs_2tap_avx2(filter_params_y, subpel_y_q4, &coeffs_256);
if (w == 8) {
__m128i s_128[2];
__m256i r[2];
s_128[0] = _mm_loadu_si128((__m128i *)im);
do {
xy_y_convolve_2tap_8x2_avx2(im, s_128, &coeffs_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 16) {
__m256i s_256[2], r[4];
s_256[0] = _mm256_loadu_si256((__m256i *)im);
do {
xy_y_convolve_2tap_16x2_avx2(im, s_256, &coeffs_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 32) {
__m256i s_256[2][2];
s_256[0][0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[0][1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
do {
xy_y_convolve_2tap_32_all_avx2(im + 32, s_256[0], s_256[1], &coeffs_256,
dst);
im += 2 * 32;
xy_y_convolve_2tap_32_all_avx2(im, s_256[1], s_256[0], &coeffs_256,
dst + dst_stride);
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 64) {
__m256i s_256[2][4];
s_256[0][0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[0][1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
s_256[0][2] = _mm256_loadu_si256((__m256i *)(im + 2 * 16));
s_256[0][3] = _mm256_loadu_si256((__m256i *)(im + 3 * 16));
do {
xy_y_convolve_2tap_32_all_avx2(im + 64, s_256[0] + 0, s_256[1] + 0,
&coeffs_256, dst);
xy_y_convolve_2tap_32_all_avx2(im + 96, s_256[0] + 2, s_256[1] + 2,
&coeffs_256, dst + 32);
im += 2 * 64;
xy_y_convolve_2tap_32_all_avx2(im, s_256[1] + 0, s_256[0] + 0,
&coeffs_256, dst + dst_stride);
xy_y_convolve_2tap_32_all_avx2(im + 32, s_256[1] + 2, s_256[0] + 2,
&coeffs_256, dst + dst_stride + 32);
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i s_256[2][8];
assert(w == 128);
load_16bit_8rows_avx2(im, 16, s_256[0]);
do {
xy_y_convolve_2tap_32_all_avx2(im + 128, s_256[0] + 0, s_256[1] + 0,
&coeffs_256, dst);
xy_y_convolve_2tap_32_all_avx2(im + 160, s_256[0] + 2, s_256[1] + 2,
&coeffs_256, dst + 1 * 32);
xy_y_convolve_2tap_32_all_avx2(im + 192, s_256[0] + 4, s_256[1] + 4,
&coeffs_256, dst + 2 * 32);
xy_y_convolve_2tap_32_all_avx2(im + 224, s_256[0] + 6, s_256[1] + 6,
&coeffs_256, dst + 3 * 32);
im += 2 * 128;
xy_y_convolve_2tap_32_all_avx2(im, s_256[1] + 0, s_256[0] + 0,
&coeffs_256, dst + dst_stride);
xy_y_convolve_2tap_32_all_avx2(im + 32, s_256[1] + 2, s_256[0] + 2,
&coeffs_256, dst + dst_stride + 1 * 32);
xy_y_convolve_2tap_32_all_avx2(im + 64, s_256[1] + 4, s_256[0] + 4,
&coeffs_256, dst + dst_stride + 2 * 32);
xy_y_convolve_2tap_32_all_avx2(im + 96, s_256[1] + 6, s_256[0] + 6,
&coeffs_256, dst + dst_stride + 3 * 32);
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
}
}
static void convolve_2d_sr_ver_2tap_half_avx2(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride) {
const int16_t *im = im_block;
int32_t y = h;
(void)filter_params_y;
(void)subpel_y_q4;
if (w == 2) {
__m128i s_32[2];
s_32[0] = _mm_cvtsi32_si128(*(int32_t *)im);
do {
const __m128i res = xy_y_convolve_2tap_2x2_half_pel_sse2(im, s_32);
const __m128i r = xy_y_round_half_pel_sse2(res);
pack_store_2x2_sse2(r, dst, dst_stride);
im += 2 * 2;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 4) {
__m128i s_64[2];
s_64[0] = _mm_loadl_epi64((__m128i *)im);
do {
const __m128i res = xy_y_convolve_2tap_4x2_half_pel_sse2(im, s_64);
const __m128i r = xy_y_round_half_pel_sse2(res);
pack_store_4x2_sse2(r, dst, dst_stride);
im += 2 * 4;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 8) {
__m128i s_128[2];
s_128[0] = _mm_loadu_si128((__m128i *)im);
do {
const __m256i res = xy_y_convolve_2tap_8x2_half_pel_avx2(im, s_128);
const __m256i r = xy_y_round_half_pel_avx2(res);
pack_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 16) {
__m256i s_256[2], r[2];
s_256[0] = _mm256_loadu_si256((__m256i *)im);
do {
xy_y_convolve_2tap_16x2_half_pel_avx2(im, s_256, r);
r[0] = xy_y_round_half_pel_avx2(r[0]);
r[1] = xy_y_round_half_pel_avx2(r[1]);
xy_y_pack_store_16x2_avx2(r[0], r[1], dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 32) {
__m256i s_256[2][2];
s_256[0][0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[0][1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
do {
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 32, s_256[0], s_256[1], dst);
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 2 * 32, s_256[1], s_256[0],
dst + dst_stride);
im += 2 * 32;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 64) {
__m256i s_256[2][4];
s_256[0][0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[0][1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
s_256[0][2] = _mm256_loadu_si256((__m256i *)(im + 2 * 16));
s_256[0][3] = _mm256_loadu_si256((__m256i *)(im + 3 * 16));
do {
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 64, s_256[0] + 0,
s_256[1] + 0, dst);
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 96, s_256[0] + 2,
s_256[1] + 2, dst + 32);
im += 2 * 64;
xy_y_convolve_2tap_half_pel_32_all_avx2(im, s_256[1] + 0, s_256[0] + 0,
dst + dst_stride);
xy_y_convolve_2tap_half_pel_32_all_avx2(
im + 32, s_256[1] + 2, s_256[0] + 2, dst + dst_stride + 32);
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i s_256[2][8];
assert(w == 128);
load_16bit_8rows_avx2(im, 16, s_256[0]);
do {
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 128, s_256[0] + 0,
s_256[1] + 0, dst);
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 160, s_256[0] + 2,
s_256[1] + 2, dst + 1 * 32);
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 192, s_256[0] + 4,
s_256[1] + 4, dst + 2 * 32);
xy_y_convolve_2tap_half_pel_32_all_avx2(im + 224, s_256[0] + 6,
s_256[1] + 6, dst + 3 * 32);
im += 2 * 128;
xy_y_convolve_2tap_half_pel_32_all_avx2(im, s_256[1] + 0, s_256[0] + 0,
dst + dst_stride);
xy_y_convolve_2tap_half_pel_32_all_avx2(
im + 32, s_256[1] + 2, s_256[0] + 2, dst + dst_stride + 1 * 32);
xy_y_convolve_2tap_half_pel_32_all_avx2(
im + 64, s_256[1] + 4, s_256[0] + 4, dst + dst_stride + 2 * 32);
xy_y_convolve_2tap_half_pel_32_all_avx2(
im + 96, s_256[1] + 6, s_256[0] + 6, dst + dst_stride + 3 * 32);
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
}
static void convolve_2d_sr_ver_4tap_avx2(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride) {
const int16_t *im = im_block;
int32_t y = h;
if (w == 2) {
__m128i coeffs_128[2], s_32[4], ss_128[2];
prepare_coeffs_4tap_sse2(filter_params_y, subpel_y_q4, coeffs_128);
s_32[0] = _mm_cvtsi32_si128(*(int32_t *)(im + 0 * 2));
s_32[1] = _mm_cvtsi32_si128(*(int32_t *)(im + 1 * 2));
s_32[2] = _mm_cvtsi32_si128(*(int32_t *)(im + 2 * 2));
const __m128i src01 = _mm_unpacklo_epi32(s_32[0], s_32[1]);
const __m128i src12 = _mm_unpacklo_epi32(s_32[1], s_32[2]);
ss_128[0] = _mm_unpacklo_epi16(src01, src12);
do {
const __m128i res =
xy_y_convolve_4tap_2x2_sse2(im, s_32, ss_128, coeffs_128);
xy_y_round_store_2x2_sse2(res, dst, dst_stride);
im += 2 * 2;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i coeffs_256[2];
prepare_coeffs_4tap_avx2(filter_params_y, subpel_y_q4, coeffs_256);
if (w == 4) {
__m128i s_64[4];
__m256i s_256[2], ss_256[2];
s_64[0] = _mm_loadl_epi64((__m128i *)(im + 0 * 4));
s_64[1] = _mm_loadl_epi64((__m128i *)(im + 1 * 4));
s_64[2] = _mm_loadl_epi64((__m128i *)(im + 2 * 4));
// Load lines a and b. Line a to lower 128, line b to upper 128
s_256[0] = _mm256_setr_m128i(s_64[0], s_64[1]);
s_256[1] = _mm256_setr_m128i(s_64[1], s_64[2]);
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
do {
const __m256i res =
xy_y_convolve_4tap_4x2_avx2(im, s_64, ss_256, coeffs_256);
xy_y_round_store_4x2_avx2(res, dst, dst_stride);
im += 2 * 4;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 8) {
__m256i s_256[4], r[2];
s_256[0] = _mm256_loadu_si256((__m256i *)(im + 0 * 8));
s_256[1] = _mm256_loadu_si256((__m256i *)(im + 1 * 8));
if (subpel_y_q4 != 8) {
__m256i ss_256[4];
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[2] = _mm256_unpackhi_epi16(s_256[0], s_256[1]);
do {
xy_y_convolve_4tap_8x2_avx2(im, ss_256, coeffs_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
do {
xy_y_convolve_4tap_8x2_half_pel_avx2(im, coeffs_256, s_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else if (w == 16) {
__m256i s_256[5];
s_256[0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
s_256[2] = _mm256_loadu_si256((__m256i *)(im + 2 * 16));
if (subpel_y_q4 != 8) {
__m256i ss_256[4], tt_256[4], r[4];
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[2] = _mm256_unpackhi_epi16(s_256[0], s_256[1]);
tt_256[0] = _mm256_unpacklo_epi16(s_256[1], s_256[2]);
tt_256[2] = _mm256_unpackhi_epi16(s_256[1], s_256[2]);
do {
xy_y_convolve_4tap_16x2_avx2(im, s_256, ss_256, tt_256, coeffs_256,
r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i r[4];
do {
xy_y_convolve_4tap_16x2_half_pelavx2(im, s_256, coeffs_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else {
/*It's a special condition for OBMC. A/c to Av1 spec 4-tap won't
support for width(w)>16, but for OBMC while predicting above block
it reduces size block to Wx(h/2), for example, if above block size
is 32x8, we get block size as 32x4 for OBMC.*/
int32_t x = 0;
assert(!(w % 32));
__m256i s_256[2][4], ss_256[2][4], tt_256[2][4], r0[4], r1[4];
do {
const int16_t *s = im + x;
uint8_t *d = dst + x;
loadu_unpack_16bit_3rows_avx2(s, w, s_256[0], ss_256[0], tt_256[0]);
loadu_unpack_16bit_3rows_avx2(s + 16, w, s_256[1], ss_256[1],
tt_256[1]);
y = h;
do {
xy_y_convolve_4tap_32x2_avx2(s, w, s_256[0], ss_256[0], tt_256[0],
coeffs_256, r0);
xy_y_convolve_4tap_32x2_avx2(s + 16, w, s_256[1], ss_256[1],
tt_256[1], coeffs_256, r1);
xy_y_round_store_32_avx2(r0 + 0, r1 + 0, d);
xy_y_round_store_32_avx2(r0 + 2, r1 + 2, d + dst_stride);
s += 2 * w;
d += 2 * dst_stride;
y -= 2;
} while (y);
x += 32;
} while (x < w);
}
}
}
static void convolve_2d_sr_ver_6tap_avx2(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride) {
const int16_t *im = im_block;
int32_t y;
if (w == 2) {
__m128i coeffs_128[3], s_32[6], ss_128[3];
prepare_coeffs_6tap_ssse3(filter_params_y, subpel_y_q4, coeffs_128);
s_32[0] = _mm_cvtsi32_si128(*(int32_t *)(im + 0 * 2));
s_32[1] = _mm_cvtsi32_si128(*(int32_t *)(im + 1 * 2));
s_32[2] = _mm_cvtsi32_si128(*(int32_t *)(im + 2 * 2));
s_32[3] = _mm_cvtsi32_si128(*(int32_t *)(im + 3 * 2));
s_32[4] = _mm_cvtsi32_si128(*(int32_t *)(im + 4 * 2));
const __m128i src01 = _mm_unpacklo_epi32(s_32[0], s_32[1]);
const __m128i src12 = _mm_unpacklo_epi32(s_32[1], s_32[2]);
const __m128i src23 = _mm_unpacklo_epi32(s_32[2], s_32[3]);
const __m128i src34 = _mm_unpacklo_epi32(s_32[3], s_32[4]);
ss_128[0] = _mm_unpacklo_epi16(src01, src12);
ss_128[1] = _mm_unpacklo_epi16(src23, src34);
y = h;
do {
const __m128i res =
xy_y_convolve_6tap_2x2_sse2(im, s_32, ss_128, coeffs_128);
xy_y_round_store_2x2_sse2(res, dst, dst_stride);
im += 2 * 2;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i coeffs_256[3];
prepare_coeffs_6tap_avx2(filter_params_y, subpel_y_q4, coeffs_256);
if (w == 4) {
__m128i s_64[6];
__m256i s_256[6], ss_256[3];
s_64[0] = _mm_loadl_epi64((__m128i *)(im + 0 * 4));
s_64[1] = _mm_loadl_epi64((__m128i *)(im + 1 * 4));
s_64[2] = _mm_loadl_epi64((__m128i *)(im + 2 * 4));
s_64[3] = _mm_loadl_epi64((__m128i *)(im + 3 * 4));
s_64[4] = _mm_loadl_epi64((__m128i *)(im + 4 * 4));
// Load lines a and b. Line a to lower 128, line b to upper 128
s_256[0] = _mm256_setr_m128i(s_64[0], s_64[1]);
s_256[1] = _mm256_setr_m128i(s_64[1], s_64[2]);
s_256[2] = _mm256_setr_m128i(s_64[2], s_64[3]);
s_256[3] = _mm256_setr_m128i(s_64[3], s_64[4]);
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[1] = _mm256_unpacklo_epi16(s_256[2], s_256[3]);
y = h;
do {
const __m256i res =
xy_y_convolve_6tap_4x2_avx2(im, s_64, ss_256, coeffs_256);
xy_y_round_store_4x2_avx2(res, dst, dst_stride);
im += 2 * 4;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 8) {
__m256i s_256[6], r[2];
s_256[0] = _mm256_loadu_si256((__m256i *)(im + 0 * 8));
s_256[1] = _mm256_loadu_si256((__m256i *)(im + 1 * 8));
s_256[2] = _mm256_loadu_si256((__m256i *)(im + 2 * 8));
s_256[3] = _mm256_loadu_si256((__m256i *)(im + 3 * 8));
y = h;
if (subpel_y_q4 != 8) {
__m256i ss_256[6];
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[1] = _mm256_unpacklo_epi16(s_256[2], s_256[3]);
ss_256[3] = _mm256_unpackhi_epi16(s_256[0], s_256[1]);
ss_256[4] = _mm256_unpackhi_epi16(s_256[2], s_256[3]);
do {
xy_y_convolve_6tap_8x2_avx2(im, ss_256, coeffs_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
do {
xy_y_convolve_6tap_8x2_half_pel_avx2(im, coeffs_256, s_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else if (w == 16) {
__m256i s_256[6];
s_256[0] = _mm256_loadu_si256((__m256i *)(im + 0 * 16));
s_256[1] = _mm256_loadu_si256((__m256i *)(im + 1 * 16));
s_256[2] = _mm256_loadu_si256((__m256i *)(im + 2 * 16));
s_256[3] = _mm256_loadu_si256((__m256i *)(im + 3 * 16));
s_256[4] = _mm256_loadu_si256((__m256i *)(im + 4 * 16));
y = h;
if (subpel_y_q4 != 8) {
__m256i ss_256[6], tt_256[6], r[4];
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[1] = _mm256_unpacklo_epi16(s_256[2], s_256[3]);
ss_256[3] = _mm256_unpackhi_epi16(s_256[0], s_256[1]);
ss_256[4] = _mm256_unpackhi_epi16(s_256[2], s_256[3]);
tt_256[0] = _mm256_unpacklo_epi16(s_256[1], s_256[2]);
tt_256[1] = _mm256_unpacklo_epi16(s_256[3], s_256[4]);
tt_256[3] = _mm256_unpackhi_epi16(s_256[1], s_256[2]);
tt_256[4] = _mm256_unpackhi_epi16(s_256[3], s_256[4]);
do {
xy_y_convolve_6tap_16x2_avx2(im, 16, s_256, ss_256, tt_256,
coeffs_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i ss_256[4], r[4];
do {
xy_y_convolve_6tap_16x2_half_pel_avx2(im, 16, s_256, ss_256,
coeffs_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else {
int32_t x = 0;
assert(!(w % 32));
__m256i s_256[2][6], ss_256[2][6], tt_256[2][6], r0[4], r1[4];
do {
const int16_t *s = im + x;
uint8_t *d = dst + x;
loadu_unpack_16bit_5rows_avx2(s, w, s_256[0], ss_256[0], tt_256[0]);
loadu_unpack_16bit_5rows_avx2(s + 16, w, s_256[1], ss_256[1],
tt_256[1]);
y = h;
do {
xy_y_convolve_6tap_16x2_avx2(s, w, s_256[0], ss_256[0], tt_256[0],
coeffs_256, r0);
xy_y_convolve_6tap_16x2_avx2(s + 16, w, s_256[1], ss_256[1],
tt_256[1], coeffs_256, r1);
xy_y_round_store_32_avx2(r0 + 0, r1 + 0, d);
xy_y_round_store_32_avx2(r0 + 2, r1 + 2, d + dst_stride);
s += 2 * w;
d += 2 * dst_stride;
y -= 2;
} while (y);
x += 32;
} while (x < w);
}
}
}
static void convolve_2d_sr_ver_8tap_avx2(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride) {
const int16_t *im = im_block;
int32_t y;
if (w == 2) {
__m128i coeffs_128[4], s_32[8], ss_128[4];
prepare_coeffs_8tap_sse2(filter_params_y, subpel_y_q4, coeffs_128);
s_32[0] = _mm_cvtsi32_si128(*(int32_t *)(im + 0 * 2));
s_32[1] = _mm_cvtsi32_si128(*(int32_t *)(im + 1 * 2));
s_32[2] = _mm_cvtsi32_si128(*(int32_t *)(im + 2 * 2));
s_32[3] = _mm_cvtsi32_si128(*(int32_t *)(im + 3 * 2));
s_32[4] = _mm_cvtsi32_si128(*(int32_t *)(im + 4 * 2));
s_32[5] = _mm_cvtsi32_si128(*(int32_t *)(im + 5 * 2));
s_32[6] = _mm_cvtsi32_si128(*(int32_t *)(im + 6 * 2));
const __m128i src01 = _mm_unpacklo_epi32(s_32[0], s_32[1]);
const __m128i src12 = _mm_unpacklo_epi32(s_32[1], s_32[2]);
const __m128i src23 = _mm_unpacklo_epi32(s_32[2], s_32[3]);
const __m128i src34 = _mm_unpacklo_epi32(s_32[3], s_32[4]);
const __m128i src45 = _mm_unpacklo_epi32(s_32[4], s_32[5]);
const __m128i src56 = _mm_unpacklo_epi32(s_32[5], s_32[6]);
ss_128[0] = _mm_unpacklo_epi16(src01, src12);
ss_128[1] = _mm_unpacklo_epi16(src23, src34);
ss_128[2] = _mm_unpacklo_epi16(src45, src56);
y = h;
do {
const __m128i res =
xy_y_convolve_8tap_2x2_sse2(im, s_32, ss_128, coeffs_128);
xy_y_round_store_2x2_sse2(res, dst, dst_stride);
im += 2 * 2;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
__m256i coeffs_256[4];
prepare_coeffs_8tap_avx2(filter_params_y, subpel_y_q4, coeffs_256);
if (w == 4) {
__m128i s_64[8];
__m256i s_256[8], ss_256[4];
s_64[0] = _mm_loadl_epi64((__m128i *)(im + 0 * 4));
s_64[1] = _mm_loadl_epi64((__m128i *)(im + 1 * 4));
s_64[2] = _mm_loadl_epi64((__m128i *)(im + 2 * 4));
s_64[3] = _mm_loadl_epi64((__m128i *)(im + 3 * 4));
s_64[4] = _mm_loadl_epi64((__m128i *)(im + 4 * 4));
s_64[5] = _mm_loadl_epi64((__m128i *)(im + 5 * 4));
s_64[6] = _mm_loadl_epi64((__m128i *)(im + 6 * 4));
// Load lines a and b. Line a to lower 128, line b to upper 128
s_256[0] = _mm256_setr_m128i(s_64[0], s_64[1]);
s_256[1] = _mm256_setr_m128i(s_64[1], s_64[2]);
s_256[2] = _mm256_setr_m128i(s_64[2], s_64[3]);
s_256[3] = _mm256_setr_m128i(s_64[3], s_64[4]);
s_256[4] = _mm256_setr_m128i(s_64[4], s_64[5]);
s_256[5] = _mm256_setr_m128i(s_64[5], s_64[6]);
ss_256[0] = _mm256_unpacklo_epi16(s_256[0], s_256[1]);
ss_256[1] = _mm256_unpacklo_epi16(s_256[2], s_256[3]);
ss_256[2] = _mm256_unpacklo_epi16(s_256[4], s_256[5]);
y = h;
do {
const __m256i res =
xy_y_convolve_8tap_4x2_avx2(im, s_64, ss_256, coeffs_256);
xy_y_round_store_4x2_avx2(res, dst, dst_stride);
im += 2 * 4;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else if (w == 8) {
__m256i s_256[8], r[2];
s_256[0] = _mm256_loadu_si256((__m256i *)(im + 0 * 8));
s_256[1] = _mm256_loadu_si256((__m256i *)(im + 1 * 8));
s_256[2] = _mm256_loadu_si256((__m256i *)(im + 2 * 8));
s_256[3] = _mm256_loadu_si256((__m256i *)(im + 3 * 8));
s_256[4] = _mm256_loadu_si256((__m256i *)(im + 4 * 8));
s_256[5] = _mm256_loadu_si256((__m256i *)(im + 5 * 8));
y = h;
if (subpel_y_q4 != 8) {
__m256i ss_256[8];
convolve_8tap_unpack_avx2(s_256, ss_256);
do {
xy_y_convolve_8tap_8x2_avx2(im, ss_256, coeffs_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
do {
xy_y_convolve_8tap_8x2_half_pel_avx2(im, coeffs_256, s_256, r);
xy_y_round_store_8x2_avx2(r, dst, dst_stride);
im += 2 * 8;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else if (w == 16) {
__m256i s_256[8], r[4];
load_16bit_7rows_avx2(im, 16, s_256);
y = h;
if (subpel_y_q4 != 8) {
__m256i ss_256[8], tt_256[8];
convolve_8tap_unpack_avx2(s_256, ss_256);
convolve_8tap_unpack_avx2(s_256 + 1, tt_256);
do {
xy_y_convolve_8tap_16x2_avx2(im, 16, coeffs_256, s_256, ss_256,
tt_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
} else {
do {
xy_y_convolve_8tap_16x2_half_pel_avx2(im, 16, coeffs_256, s_256, r);
xy_y_round_store_16x2_avx2(r, dst, dst_stride);
im += 2 * 16;
dst += 2 * dst_stride;
y -= 2;
} while (y);
}
} else {
int32_t x = 0;
__m256i s_256[2][8], r0[4], r1[4];
assert(!(w % 32));
__m256i ss_256[2][8], tt_256[2][8];
do {
const int16_t *s = im + x;
uint8_t *d = dst + x;
load_16bit_7rows_avx2(s, w, s_256[0]);
convolve_8tap_unpack_avx2(s_256[0], ss_256[0]);
convolve_8tap_unpack_avx2(s_256[0] + 1, tt_256[0]);
load_16bit_7rows_avx2(s + 16, w, s_256[1]);
convolve_8tap_unpack_avx2(s_256[1], ss_256[1]);
convolve_8tap_unpack_avx2(s_256[1] + 1, tt_256[1]);
y = h;
do {
xy_y_convolve_8tap_16x2_avx2(s, w, coeffs_256, s_256[0], ss_256[0],
tt_256[0], r0);
xy_y_convolve_8tap_16x2_avx2(s + 16, w, coeffs_256, s_256[1],
ss_256[1], tt_256[1], r1);
xy_y_round_store_32_avx2(r0 + 0, r1 + 0, d);
xy_y_round_store_32_avx2(r0 + 2, r1 + 2, d + dst_stride);
s += 2 * w;
d += 2 * dst_stride;
y -= 2;
} while (y);
x += 32;
} while (x < w);
}
}
}
typedef void (*Convolve2dSrHorTapFunc)(
const uint8_t *const src, const int32_t src_stride, const int32_t w,
const int32_t h, const InterpFilterParams *const filter_params_x,
const int32_t subpel_x_q4, int16_t *const im_block);
typedef void (*Convolve2dSrVerTapFunc)(
const int16_t *const im_block, const int32_t w, const int32_t h,
const InterpFilterParams *const filter_params_y, const int32_t subpel_y_q4,
uint8_t *dst, const int32_t dst_stride);
static AOM_FORCE_INLINE void av1_convolve_2d_sr_specialized_avx2(
const uint8_t *src, int32_t src_stride, uint8_t *dst, int32_t dst_stride,
int32_t w, int32_t h, const InterpFilterParams *filter_params_x,
const InterpFilterParams *filter_params_y, const int32_t subpel_x_q4,
const int32_t subpel_y_q4, ConvolveParams *conv_params) {
static const Convolve2dSrHorTapFunc
convolve_2d_sr_hor_tap_func_table[MAX_FILTER_TAP + 1] = {
NULL,
NULL,
convolve_2d_sr_hor_2tap_avx2,
NULL,
convolve_2d_sr_hor_4tap_ssse3,
NULL,
convolve_2d_sr_hor_6tap_avx2,
NULL,
convolve_2d_sr_hor_8tap_avx2
};
static const Convolve2dSrVerTapFunc
convolve_2d_sr_ver_tap_func_table[MAX_FILTER_TAP + 1] = {
NULL,
convolve_2d_sr_ver_2tap_half_avx2,
convolve_2d_sr_ver_2tap_avx2,
convolve_2d_sr_ver_4tap_avx2,
convolve_2d_sr_ver_4tap_avx2,
convolve_2d_sr_ver_6tap_avx2,
convolve_2d_sr_ver_6tap_avx2,
convolve_2d_sr_ver_8tap_avx2,
convolve_2d_sr_ver_8tap_avx2
};
const int32_t tap_x = get_filter_tap(filter_params_x, subpel_x_q4);
const int32_t tap_y = get_filter_tap(filter_params_y, subpel_y_q4);
assert(tap_x != 12 && tap_y != 12);
const uint8_t *src_ptr = src - ((tap_y >> 1) - 1) * src_stride;
// Note: im_block is 8-pixel interlaced for width 32 and up, to avoid data
// permutation.
DECLARE_ALIGNED(32, int16_t,
im_block[(MAX_SB_SIZE + MAX_FILTER_TAP) * MAX_SB_SIZE]);
(void)conv_params;
assert(conv_params->round_0 == 3);
assert(conv_params->round_1 == 11);
// horizontal filter
int32_t hh = h + tap_y;
assert(!(hh % 2));
convolve_2d_sr_hor_tap_func_table[tap_x](
src_ptr, src_stride, w, hh, filter_params_x, subpel_x_q4, im_block);
// vertical filter
convolve_2d_sr_ver_tap_func_table[tap_y - (subpel_y_q4 == 8)](
im_block, w, h, filter_params_y, subpel_y_q4, dst, dst_stride);
}
#endif // THIRD_PARTY_SVT_AV1_CONVOLVE_2D_AVX2_H_