/* * Copyright (c) 2018, 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_AVX2_H_ #define THIRD_PARTY_SVT_AV1_CONVOLVE_AVX2_H_ #include "EbMemory_AVX2.h" #include "EbMemory_SSE4_1.h" #include "synonyms.h" #include "aom_dsp/aom_filter.h" #include "aom_dsp/x86/convolve_avx2.h" #include "aom_dsp/x86/mem_sse2.h" static inline void populate_coeffs_4tap_avx2(const __m128i coeffs_128, __m256i coeffs[2]) { … } static inline void populate_coeffs_6tap_avx2(const __m128i coeffs_128, __m256i coeffs[3]) { … } static inline void populate_coeffs_8tap_avx2(const __m128i coeffs_128, __m256i coeffs[4]) { … } static inline void prepare_half_coeffs_2tap_ssse3( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [1] */) { … } static inline void prepare_half_coeffs_4tap_ssse3( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [2] */) { … } static inline void prepare_half_coeffs_6tap_ssse3( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [3] */) { … } static inline void prepare_half_coeffs_8tap_ssse3( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [4] */) { … } static inline void prepare_half_coeffs_2tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [1] */) { … } static inline void prepare_half_coeffs_4tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [2] */) { … } static inline void prepare_half_coeffs_6tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [3] */) { … } static inline void prepare_half_coeffs_8tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [4] */) { … } static inline void prepare_coeffs_2tap_sse2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [1] */) { … } static inline void prepare_coeffs_4tap_sse2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [2] */) { … } static inline void prepare_coeffs_6tap_ssse3( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [3] */) { … } static inline void prepare_coeffs_8tap_sse2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m128i *const coeffs /* [4] */) { … } static inline void prepare_coeffs_2tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [1] */) { … } static inline void prepare_coeffs_4tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [2] */) { … } static inline void prepare_coeffs_6tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [3]*/) { … } static inline void prepare_coeffs_8tap_avx2( const InterpFilterParams *const filter_params, const int32_t subpel_q4, __m256i *const coeffs /* [4] */) { … } static inline void load_16bit_5rows_avx2(const int16_t *const src, const ptrdiff_t stride, __m256i dst[5]) { … } static inline void load_16bit_7rows_avx2(const int16_t *const src, const ptrdiff_t stride, __m256i dst[7]) { … } static AOM_FORCE_INLINE void load_16bit_8rows_avx2(const int16_t *const src, const ptrdiff_t stride, __m256i dst[8]) { … } static AOM_FORCE_INLINE void loadu_unpack_16bit_5rows_avx2( const int16_t *const src, const ptrdiff_t stride, __m256i s_256[5], __m256i ss_256[5], __m256i tt_256[5]) { … } static AOM_FORCE_INLINE void loadu_unpack_16bit_3rows_avx2( const int16_t *const src, const ptrdiff_t stride, __m256i s_256[3], __m256i ss_256[3], __m256i tt_256[3]) { … } static inline void convolve_8tap_unpack_avx2(const __m256i s[6], __m256i ss[7]) { … } static inline __m128i convolve_2tap_ssse3(const __m128i ss[1], const __m128i coeffs[1]) { … } static inline __m128i convolve_4tap_ssse3(const __m128i ss[2], const __m128i coeffs[2]) { … } static inline __m128i convolve_6tap_ssse3(const __m128i ss[3], const __m128i coeffs[3]) { … } static inline __m128i convolve_8tap_ssse3(const __m128i ss[4], const __m128i coeffs[4]) { … } static inline __m256i convolve_2tap_avx2(const __m256i ss[1], const __m256i coeffs[1]) { … } static inline __m256i convolve_4tap_avx2(const __m256i ss[2], const __m256i coeffs[2]) { … } static inline __m256i convolve_6tap_avx2(const __m256i ss[3], const __m256i coeffs[3]) { … } static inline __m256i convolve_8tap_avx2(const __m256i ss[4], const __m256i coeffs[4]) { … } static inline __m128i convolve16_2tap_sse2(const __m128i ss[1], const __m128i coeffs[1]) { … } static inline __m128i convolve16_4tap_sse2(const __m128i ss[2], const __m128i coeffs[2]) { … } static inline __m128i convolve16_6tap_sse2(const __m128i ss[3], const __m128i coeffs[3]) { … } static inline __m128i convolve16_8tap_sse2(const __m128i ss[4], const __m128i coeffs[4]) { … } static inline __m256i convolve16_2tap_avx2(const __m256i ss[1], const __m256i coeffs[1]) { … } static inline __m256i convolve16_4tap_avx2(const __m256i ss[2], const __m256i coeffs[2]) { … } static inline __m256i convolve16_6tap_avx2(const __m256i ss[3], const __m256i coeffs[3]) { … } static inline __m256i convolve16_8tap_avx2(const __m256i ss[4], const __m256i coeffs[4]) { … } static inline __m256i x_convolve_4tap_avx2(const __m256i data, const __m256i coeffs[2], const __m256i filt[2]) { … } static inline __m256i x_convolve_6tap_avx2(const __m256i data, const __m256i coeffs[3], const __m256i filt[3]) { … } static inline __m256i x_convolve_8tap_avx2(const __m256i data, const __m256i coeffs[4], const __m256i filt[4]) { … } static inline __m256i sr_y_round_avx2(const __m256i src) { … } static inline __m128i xy_x_round_sse2(const __m128i src) { … } static inline __m256i xy_x_round_avx2(const __m256i src) { … } static inline void xy_x_round_store_2x2_sse2(const __m128i res, int16_t *const dst) { … } static inline void xy_x_round_store_4x2_sse2(const __m128i res, int16_t *const dst) { … } static inline void xy_x_round_store_8x2_sse2(const __m128i res[2], int16_t *const dst) { … } static inline void xy_x_round_store_8x2_avx2(const __m256i res, int16_t *const dst) { … } static inline void xy_x_round_store_32_avx2(const __m256i res[2], int16_t *const dst) { … } static inline __m128i xy_y_round_sse2(const __m128i src) { … } static inline __m128i xy_y_round_half_pel_sse2(const __m128i src) { … } static inline __m256i xy_y_round_avx2(const __m256i src) { … } static inline __m256i xy_y_round_16_avx2(const __m256i r[2]) { … } static inline __m256i xy_y_round_half_pel_avx2(const __m256i src) { … } static inline void pack_store_2x2_sse2(const __m128i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void pack_store_4x2_sse2(const __m128i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void pack_store_4x2_avx2(const __m256i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void pack_store_8x2_avx2(const __m256i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void pack_store_16x2_avx2(const __m256i res0, const __m256i res1, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void xy_y_pack_store_16x2_avx2(const __m256i res0, const __m256i res1, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void pack_store_32_avx2(const __m256i res0, const __m256i res1, uint8_t *const dst) { … } static inline void xy_y_round_store_2x2_sse2(const __m128i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void xy_y_round_store_4x2_avx2(const __m256i res, uint8_t *const dst, const ptrdiff_t stride) { … } static inline void xy_y_pack_store_32_avx2(const __m256i res0, const __m256i res1, uint8_t *const dst) { … } static inline void xy_y_round_store_32_avx2(const __m256i r0[2], const __m256i r1[2], uint8_t *const dst) { … } static inline void convolve_store_32_avx2(const __m256i res0, const __m256i res1, uint8_t *const dst) { … } static inline __m128i sr_x_round_sse2(const __m128i src) { … } static inline __m256i sr_x_round_avx2(const __m256i src) { … } static inline __m128i sr_y_round_sse2(const __m128i src) { … } static inline void sr_x_round_store_8x2_avx2(const __m256i res, uint8_t *const dst, const ptrdiff_t dst_stride) { … } static inline void sr_x_round_store_16x2_avx2(const __m256i res[2], uint8_t *const dst, const ptrdiff_t dst_stride) { … } static inline void sr_x_round_store_32_avx2(const __m256i res[2], uint8_t *const dst) { … } static inline void sr_y_round_store_8x2_avx2(const __m256i res, uint8_t *const dst, const ptrdiff_t dst_stride) { … } static inline void sr_y_round_store_16x2_avx2(const __m256i res[2], uint8_t *const dst, const ptrdiff_t dst_stride) { … } static inline void sr_y_2tap_32_avg_avx2(const uint8_t *const src, const __m256i s0, __m256i *const s1, uint8_t *const dst) { … } static inline void sr_x_2tap_32_avg_avx2(const uint8_t *const src, uint8_t *const dst) { … } static inline __m128i x_convolve_2tap_2x2_sse4_1(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[1]) { … } static inline __m128i x_convolve_2tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[1]) { … } static inline void x_convolve_2tap_8x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[1], __m128i r[2]) { … } static inline __m256i x_convolve_2tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[1]) { … } static inline void x_convolve_2tap_16x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[1], __m256i r[2]) { … } static inline void x_convolve_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], __m256i r[2]) { … } static inline __m128i x_convolve_4tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[2]) { … } static inline __m128i x_convolve_4tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[2]) { … } static inline __m256i x_convolve_4tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[2], const __m256i filt[2]) { … } static inline void x_convolve_4tap_16x2_avx2(const uint8_t *const src, const int32_t src_stride, const __m256i coeffs[2], const __m256i filt[2], __m256i r[2]) { … } static inline void x_convolve_4tap_32_avx2(const uint8_t *const src, const __m256i coeffs[2], const __m256i filt[2], __m256i r[2]) { … } static inline __m128i x_convolve_6tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[3]) { … } static inline __m128i x_convolve_6tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[3]) { … } static inline __m256i x_convolve_6tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[3], const __m256i filt[3]) { … } static inline void x_convolve_6tap_16x2_avx2(const uint8_t *const src, const int32_t src_stride, const __m256i coeffs[3], const __m256i filt[3], __m256i r[2]) { … } static inline void x_convolve_6tap_32_avx2(const uint8_t *const src, const __m256i coeffs[3], const __m256i filt[3], __m256i r[2]) { … } static inline __m256i x_convolve_8tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], const __m256i filt[4]) { … } static AOM_FORCE_INLINE void x_convolve_8tap_16x2_avx2(const uint8_t *const src, const int32_t src_stride, const __m256i coeffs[4], const __m256i filt[4], __m256i r[2]) { … } static AOM_FORCE_INLINE void x_convolve_8tap_32_avx2(const uint8_t *const src, const __m256i coeffs[4], const __m256i filt[4], __m256i r[2]) { … } static inline __m128i y_convolve_2tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[1], __m128i s_16[2]) { … } static inline __m128i y_convolve_2tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[1], __m128i s_32[2]) { … } static inline __m256i y_convolve_2tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[1], __m128i s_64[2]) { … } static inline void y_convolve_2tap_16x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[1], __m128i s_128[2], __m256i r[2]) { … } static inline void y_convolve_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], const __m256i s0, __m256i *const s1, __m256i r[2]) { … } static inline __m128i y_convolve_4tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[2], __m128i s_16[4], __m128i ss_128[2]) { … } static inline __m128i y_convolve_4tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[2], __m128i s_32[4], __m128i ss_128[2]) { … } static inline __m256i y_convolve_4tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[2], __m128i s_64[4], __m256i ss_256[2]) { … } static inline void y_convolve_4tap_16x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[2], __m128i s_128[4], __m256i ss_256[4], __m256i r[2]) { … } static inline __m128i y_convolve_6tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[3], __m128i s_16[6], __m128i ss_128[3]) { … } static inline void y_convolve_4tap_32x2_avx2( const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[2], __m256i s_256[4], __m256i ss_256[4], __m256i tt_256[4], __m256i r[4]) { … } static inline __m128i y_convolve_6tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[3], __m128i s_32[6], __m128i ss_128[3]) { … } static inline __m256i y_convolve_6tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[3], __m128i s_64[6], __m256i ss_256[3]) { … } static inline void y_convolve_6tap_16x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[3], __m128i s_128[6], __m256i ss_256[6], __m256i r[2]) { … } static inline void y_convolve_6tap_32x2_avx2( const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[3], __m256i s_256[6], __m256i ss_256[6], __m256i tt_256[6], __m256i r[4]) { … } static inline __m128i y_convolve_8tap_2x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[4], __m128i s_16[8], __m128i ss_128[4]) { … } static inline __m128i y_convolve_8tap_4x2_ssse3(const uint8_t *const src, const ptrdiff_t stride, const __m128i coeffs[4], __m128i s_32[8], __m128i ss_128[4]) { … } static inline __m256i y_convolve_8tap_8x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], __m128i s_64[8], __m256i ss_256[4]) { … } static inline void y_convolve_8tap_16x2_avx2(const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], __m128i s_128[8], __m256i ss_256[8], __m256i r[2]) { … } static inline void y_convolve_8tap_32x2_avx2( const uint8_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], __m256i s_256[8], __m256i ss_256[8], __m256i tt_256[8], __m256i r[4]) { … } static inline void xy_x_convolve_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], __m256i r[2]) { … } static inline void xy_x_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], int16_t *const dst) { … } static inline void xy_x_4tap_32_avx2(const uint8_t *const src, const __m256i coeffs[2], const __m256i filt[2], int16_t *const dst) { … } static inline void xy_x_6tap_32_avx2(const uint8_t *const src, const __m256i coeffs[3], const __m256i filt[3], int16_t *const dst) { … } static inline void xy_x_8tap_32_avx2(const uint8_t *const src, const __m256i coeffs[4], const __m256i filt[4], int16_t *const dst) { … } static inline __m128i xy_y_convolve_2tap_2x2_sse2(const int16_t *const src, __m128i s_32[2], const __m128i coeffs[1]) { … } static inline __m128i xy_y_convolve_2tap_2x2_half_pel_sse2( const int16_t *const src, __m128i s_32[2]) { … } static inline void xy_y_convolve_2tap_4x2_sse2(const int16_t *const src, __m128i s_64[2], const __m128i coeffs[1], __m128i r[2]) { … } static inline __m128i xy_y_convolve_2tap_4x2_half_pel_sse2( const int16_t *const src, __m128i s_64[2]) { … } static inline void xy_y_convolve_2tap_16_avx2(const __m256i s0, const __m256i s1, const __m256i coeffs[1], __m256i r[2]) { … } static inline void xy_y_convolve_2tap_8x2_avx2(const int16_t *const src, __m128i s_128[2], const __m256i coeffs[1], __m256i r[2]) { … } static inline __m256i xy_y_convolve_2tap_8x2_half_pel_avx2( const int16_t *const src, __m128i s_128[2]) { … } static inline void xy_y_convolve_2tap_16x2_half_pel_avx2( const int16_t *const src, __m256i s_256[2], __m256i r[2]) { … } static inline void xy_y_store_16x2_avx2(const __m256i r[2], uint8_t *const dst, const ptrdiff_t stride) { … } static inline void xy_y_convolve_2tap_16x2_avx2(const int16_t *const src, __m256i s[2], const __m256i coeffs[1], __m256i r[4]) { … } static inline void xy_y_convolve_2tap_32_avx2(const int16_t *const src, const __m256i s0[2], __m256i s1[2], const __m256i coeffs[1], __m256i r[4]) { … } static inline void xy_y_convolve_2tap_32_all_avx2(const int16_t *const src, const __m256i s0[2], __m256i s1[2], const __m256i coeffs[1], uint8_t *const dst) { … } static inline void xy_y_convolve_2tap_half_pel_32_avx2(const int16_t *const src, const __m256i s0[2], __m256i s1[2], __m256i r[2]) { … } static inline void xy_y_convolve_2tap_half_pel_32_all_avx2( const int16_t *const src, const __m256i s0[2], __m256i s1[2], uint8_t *const dst) { … } static inline __m128i xy_y_convolve_4tap_2x2_sse2(const int16_t *const src, __m128i s_32[4], __m128i ss_128[2], const __m128i coeffs[2]) { … } static inline __m256i xy_y_convolve_4tap_4x2_avx2(const int16_t *const src, __m128i s_64[4], __m256i ss_256[2], const __m256i coeffs[2]) { … } static inline void xy_y_convolve_4tap_16_avx2(const __m256i *const ss, const __m256i coeffs[2], __m256i r[2]) { … } static inline void xy_y_convolve_4tap_8x2_avx2(const int16_t *const src, __m256i ss_256[4], const __m256i coeffs[2], __m256i r[2]) { … } static inline void xy_y_convolve_4tap_8x2_half_pel_avx2( const int16_t *const src, const __m256i coeffs[1], __m256i s_256[4], __m256i r[2]) { … } static inline void xy_y_convolve_4tap_16x2_avx2( const int16_t *const src, __m256i s_256[4], __m256i ss_256[4], __m256i tt_256[4], const __m256i coeffs[2], __m256i r[4]) { … } static inline void xy_y_convolve_4tap_32x2_avx2( const int16_t *const src, const ptrdiff_t stride, __m256i s_256[4], __m256i ss_256[4], __m256i tt_256[4], const __m256i coeffs[2], __m256i r[4]) { … } static inline void xy_y_convolve_4tap_16x2_half_pelavx2( const int16_t *const src, __m256i s_256[5], const __m256i coeffs[1], __m256i r[4]) { … } static inline __m128i xy_y_convolve_6tap_2x2_sse2(const int16_t *const src, __m128i s_32[6], __m128i ss_128[3], const __m128i coeffs[3]) { … } static inline __m256i xy_y_convolve_6tap_4x2_avx2(const int16_t *const src, __m128i s_64[6], __m256i ss_256[3], const __m256i coeffs[3]) { … } static inline void xy_y_convolve_6tap_16_avx2(const __m256i ss[6], const __m256i coeffs[3], __m256i r[2]) { … } static inline void xy_y_convolve_6tap_8x2_avx2(const int16_t *const src, __m256i ss_256[6], const __m256i coeffs[3], __m256i r[2]) { … } static inline void xy_y_convolve_6tap_8x2_half_pel_avx2( const int16_t *const src, const __m256i coeffs[2], __m256i s_256[6], __m256i r[2]) { … } static inline void xy_y_convolve_6tap_16x2_avx2( const int16_t *const src, const ptrdiff_t stride, __m256i s_256[6], __m256i ss_256[6], __m256i tt_256[6], const __m256i coeffs[3], __m256i r[4]) { … } static inline void xy_y_convolve_6tap_16x2_half_pel_avx2( const int16_t *const src, const ptrdiff_t stride, __m256i s_256[6], __m256i ss_256[4], const __m256i coeffs[2], __m256i r[4]) { … } static inline __m128i xy_y_convolve_8tap_2x2_sse2(const int16_t *const src, __m128i s_32[8], __m128i ss_128[4], const __m128i coeffs[4]) { … } static inline __m256i xy_y_convolve_8tap_4x2_avx2(const int16_t *const src, __m128i s_64[8], __m256i ss_256[4], const __m256i coeffs[4]) { … } static inline void xy_y_convolve_8tap_16_avx2(const __m256i *const ss, const __m256i coeffs[4], __m256i r[2]) { … } static inline void xy_y_convolve_8tap_8x2_avx2(const int16_t *const src, __m256i ss_256[8], const __m256i coeffs[4], __m256i r[2]) { … } static inline void xy_y_convolve_8tap_8x2_half_pel_avx2( const int16_t *const src, const __m256i coeffs[2], __m256i s_256[8], __m256i r[2]) { … } static AOM_FORCE_INLINE void xy_y_convolve_8tap_16x2_avx2( const int16_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], __m256i s_256[8], __m256i ss_256[8], __m256i tt_256[8], __m256i r[4]) { … } static inline void xy_y_convolve_8tap_16x2_half_pel_avx2( const int16_t *const src, const ptrdiff_t stride, const __m256i coeffs[4], __m256i s_256[8], __m256i r[4]) { … } static inline void xy_y_round_store_8x2_avx2(const __m256i res[2], uint8_t *const dst, const ptrdiff_t stride) { … } static inline void xy_y_round_store_16x2_avx2(const __m256i res[4], uint8_t *const dst, const ptrdiff_t stride) { … } static inline void sr_y_round_store_32_avx2(const __m256i res[2], uint8_t *const dst) { … } static inline void sr_y_round_store_32x2_avx2(const __m256i res[4], uint8_t *const dst, const int32_t dst_stride) { … } static inline void sr_y_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], const __m256i s0, __m256i *const s1, uint8_t *const dst) { … } static AOM_FORCE_INLINE void av1_convolve_y_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_y, const int32_t subpel_y_q4) { … } static inline void sr_x_2tap_32_avx2(const uint8_t *const src, const __m256i coeffs[1], uint8_t *const dst) { … } static inline void sr_x_6tap_32_avx2(const uint8_t *const src, const __m256i coeffs[3], const __m256i filt[3], uint8_t *const dst) { … } static AOM_FORCE_INLINE void sr_x_8tap_32_avx2(const uint8_t *const src, const __m256i coeffs[4], const __m256i filt[4], uint8_t *const dst) { … } static AOM_FORCE_INLINE void av1_convolve_x_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 int32_t subpel_x_q4, ConvolveParams *conv_params) { … } #endif // THIRD_PARTY_SVT_AV1_CONVOLVE_AVX2_H_
Codebase Browser
chromium