/src/aom/av1/common/cfl.h
Line | Count | Source |
1 | | /* |
2 | | * Copyright (c) 2016, Alliance for Open Media. All rights reserved. |
3 | | * |
4 | | * This source code is subject to the terms of the BSD 2 Clause License and |
5 | | * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License |
6 | | * was not distributed with this source code in the LICENSE file, you can |
7 | | * obtain it at www.aomedia.org/license/software. If the Alliance for Open |
8 | | * Media Patent License 1.0 was not distributed with this source code in the |
9 | | * PATENTS file, you can obtain it at www.aomedia.org/license/patent. |
10 | | */ |
11 | | |
12 | | #ifndef AOM_AV1_COMMON_CFL_H_ |
13 | | #define AOM_AV1_COMMON_CFL_H_ |
14 | | |
15 | | #include "av1/common/av1_common_int.h" |
16 | | #include "av1/common/blockd.h" |
17 | | |
18 | | // Can we use CfL for the current block? |
19 | 10.8M | static inline CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { |
20 | 10.8M | const MB_MODE_INFO *mbmi = xd->mi[0]; |
21 | 10.8M | const BLOCK_SIZE bsize = mbmi->bsize; |
22 | 10.8M | assert(bsize < BLOCK_SIZES_ALL); |
23 | 10.8M | if (xd->lossless[mbmi->segment_id]) { |
24 | | // In lossless, CfL is available when the partition size is equal to the |
25 | | // transform size. |
26 | 51.1k | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; |
27 | 51.1k | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; |
28 | 51.1k | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); |
29 | 51.1k | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); |
30 | 51.1k | } |
31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 |
32 | 10.8M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && |
33 | 10.0M | block_size_high[bsize] <= 32); |
34 | 10.8M | } Unexecuted instantiation: decodeframe.c:is_cfl_allowed decodemv.c:is_cfl_allowed Line | Count | Source | 19 | 8.23M | static inline CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { | 20 | 8.23M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 21 | 8.23M | const BLOCK_SIZE bsize = mbmi->bsize; | 22 | 8.23M | assert(bsize < BLOCK_SIZES_ALL); | 23 | 8.23M | if (xd->lossless[mbmi->segment_id]) { | 24 | | // In lossless, CfL is available when the partition size is equal to the | 25 | | // transform size. | 26 | 50.7k | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; | 27 | 50.7k | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; | 28 | 50.7k | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); | 29 | 50.7k | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); | 30 | 50.7k | } | 31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 | 32 | 8.18M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && | 33 | 7.38M | block_size_high[bsize] <= 32); | 34 | 8.23M | } |
Line | Count | Source | 19 | 2.65M | static inline CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { | 20 | 2.65M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 21 | 2.65M | const BLOCK_SIZE bsize = mbmi->bsize; | 22 | 2.65M | assert(bsize < BLOCK_SIZES_ALL); | 23 | 2.65M | if (xd->lossless[mbmi->segment_id]) { | 24 | | // In lossless, CfL is available when the partition size is equal to the | 25 | | // transform size. | 26 | 446 | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; | 27 | 446 | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; | 28 | 446 | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); | 29 | 446 | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); | 30 | 446 | } | 31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 | 32 | 2.65M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && | 33 | 2.65M | block_size_high[bsize] <= 32); | 34 | 2.65M | } |
Unexecuted instantiation: reconintra.c:is_cfl_allowed Unexecuted instantiation: cfl_sse2.c:is_cfl_allowed Unexecuted instantiation: cfl_ssse3.c:is_cfl_allowed Unexecuted instantiation: cfl_avx2.c:is_cfl_allowed |
35 | | |
36 | | // Do we need to save the luma pixels from the current block, |
37 | | // for a possible future CfL prediction? |
38 | | static inline CFL_ALLOWED_TYPE store_cfl_required(const AV1_COMMON *cm, |
39 | 47.4M | const MACROBLOCKD *xd) { |
40 | 47.4M | const MB_MODE_INFO *mbmi = xd->mi[0]; |
41 | | |
42 | 47.4M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; |
43 | | |
44 | 45.5M | if (!xd->is_chroma_ref) { |
45 | | // For non-chroma-reference blocks, we should always store the luma pixels, |
46 | | // in case the corresponding chroma-reference block uses CfL. |
47 | | // Note that this can only happen for block sizes which are <8 on |
48 | | // their shortest side, as otherwise they would be chroma reference |
49 | | // blocks. |
50 | 2.29M | return CFL_ALLOWED; |
51 | 2.29M | } |
52 | | |
53 | | // If this block has chroma information, we know whether we're |
54 | | // actually going to perform a CfL prediction |
55 | 43.2M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && |
56 | 37.3M | mbmi->uv_mode == UV_CFL_PRED); |
57 | 45.5M | } decodeframe.c:store_cfl_required Line | Count | Source | 39 | 34.2M | const MACROBLOCKD *xd) { | 40 | 34.2M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 41 | | | 42 | 34.2M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; | 43 | | | 44 | 33.0M | if (!xd->is_chroma_ref) { | 45 | | // For non-chroma-reference blocks, we should always store the luma pixels, | 46 | | // in case the corresponding chroma-reference block uses CfL. | 47 | | // Note that this can only happen for block sizes which are <8 on | 48 | | // their shortest side, as otherwise they would be chroma reference | 49 | | // blocks. | 50 | 1.12M | return CFL_ALLOWED; | 51 | 1.12M | } | 52 | | | 53 | | // If this block has chroma information, we know whether we're | 54 | | // actually going to perform a CfL prediction | 55 | 31.9M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && | 56 | 29.1M | mbmi->uv_mode == UV_CFL_PRED); | 57 | 33.0M | } |
decodemv.c:store_cfl_required Line | Count | Source | 39 | 13.2M | const MACROBLOCKD *xd) { | 40 | 13.2M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 41 | | | 42 | 13.2M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; | 43 | | | 44 | 12.5M | if (!xd->is_chroma_ref) { | 45 | | // For non-chroma-reference blocks, we should always store the luma pixels, | 46 | | // in case the corresponding chroma-reference block uses CfL. | 47 | | // Note that this can only happen for block sizes which are <8 on | 48 | | // their shortest side, as otherwise they would be chroma reference | 49 | | // blocks. | 50 | 1.17M | return CFL_ALLOWED; | 51 | 1.17M | } | 52 | | | 53 | | // If this block has chroma information, we know whether we're | 54 | | // actually going to perform a CfL prediction | 55 | 11.3M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && | 56 | 8.23M | mbmi->uv_mode == UV_CFL_PRED); | 57 | 12.5M | } |
Unexecuted instantiation: cfl.c:store_cfl_required Unexecuted instantiation: reconintra.c:store_cfl_required Unexecuted instantiation: cfl_sse2.c:store_cfl_required Unexecuted instantiation: cfl_ssse3.c:store_cfl_required Unexecuted instantiation: cfl_avx2.c:store_cfl_required |
58 | | |
59 | 0 | static inline int get_scaled_luma_q0(int alpha_q3, int16_t pred_buf_q3) { |
60 | 0 | int scaled_luma_q6 = alpha_q3 * pred_buf_q3; |
61 | 0 | return ROUND_POWER_OF_TWO_SIGNED(scaled_luma_q6, 6); |
62 | 0 | } Unexecuted instantiation: decodeframe.c:get_scaled_luma_q0 Unexecuted instantiation: decodemv.c:get_scaled_luma_q0 Unexecuted instantiation: cfl.c:get_scaled_luma_q0 Unexecuted instantiation: reconintra.c:get_scaled_luma_q0 Unexecuted instantiation: cfl_sse2.c:get_scaled_luma_q0 Unexecuted instantiation: cfl_ssse3.c:get_scaled_luma_q0 Unexecuted instantiation: cfl_avx2.c:get_scaled_luma_q0 |
63 | | |
64 | 2.65M | static inline CFL_PRED_TYPE get_cfl_pred_type(int plane) { |
65 | 2.65M | assert(plane > 0); |
66 | 2.65M | return (CFL_PRED_TYPE)(plane - 1); |
67 | 2.65M | } Unexecuted instantiation: decodeframe.c:get_cfl_pred_type Unexecuted instantiation: decodemv.c:get_cfl_pred_type Unexecuted instantiation: cfl.c:get_cfl_pred_type reconintra.c:get_cfl_pred_type Line | Count | Source | 64 | 2.65M | static inline CFL_PRED_TYPE get_cfl_pred_type(int plane) { | 65 | 2.65M | assert(plane > 0); | 66 | 2.65M | return (CFL_PRED_TYPE)(plane - 1); | 67 | 2.65M | } |
Unexecuted instantiation: cfl_sse2.c:get_cfl_pred_type Unexecuted instantiation: cfl_ssse3.c:get_cfl_pred_type Unexecuted instantiation: cfl_avx2.c:get_cfl_pred_type |
68 | | |
69 | 280k | static inline void clear_cfl_dc_pred_cache_flags(CFL_CTX *cfl) { |
70 | 280k | cfl->use_dc_pred_cache = false; |
71 | 280k | cfl->dc_pred_is_cached[CFL_PRED_U] = false; |
72 | 280k | cfl->dc_pred_is_cached[CFL_PRED_V] = false; |
73 | 280k | } Unexecuted instantiation: decodeframe.c:clear_cfl_dc_pred_cache_flags Unexecuted instantiation: decodemv.c:clear_cfl_dc_pred_cache_flags cfl.c:clear_cfl_dc_pred_cache_flags Line | Count | Source | 69 | 280k | static inline void clear_cfl_dc_pred_cache_flags(CFL_CTX *cfl) { | 70 | 280k | cfl->use_dc_pred_cache = false; | 71 | 280k | cfl->dc_pred_is_cached[CFL_PRED_U] = false; | 72 | | cfl->dc_pred_is_cached[CFL_PRED_V] = false; | 73 | 280k | } |
Unexecuted instantiation: reconintra.c:clear_cfl_dc_pred_cache_flags Unexecuted instantiation: cfl_sse2.c:clear_cfl_dc_pred_cache_flags Unexecuted instantiation: cfl_ssse3.c:clear_cfl_dc_pred_cache_flags Unexecuted instantiation: cfl_avx2.c:clear_cfl_dc_pred_cache_flags |
74 | | |
75 | | void av1_cfl_predict_block(MACROBLOCKD *const xd, uint8_t *dst, int dst_stride, |
76 | | TX_SIZE tx_size, int plane); |
77 | | |
78 | | void cfl_store_block(MACROBLOCKD *const xd, BLOCK_SIZE bsize, TX_SIZE tx_size); |
79 | | |
80 | | void cfl_store_tx(MACROBLOCKD *const xd, int row, int col, TX_SIZE tx_size, |
81 | | BLOCK_SIZE bsize); |
82 | | |
83 | | void cfl_store_dc_pred(MACROBLOCKD *const xd, const uint8_t *input, |
84 | | CFL_PRED_TYPE pred_plane, int width); |
85 | | |
86 | | void cfl_load_dc_pred(MACROBLOCKD *const xd, uint8_t *dst, int dst_stride, |
87 | | TX_SIZE tx_size, CFL_PRED_TYPE pred_plane); |
88 | | |
89 | | // Allows the CFL_SUBSAMPLE function to switch types depending on the bitdepth. |
90 | | #define CFL_lbd_TYPE uint8_t *cfl_type |
91 | | #define CFL_hbd_TYPE uint16_t *cfl_type |
92 | | |
93 | | // Declare a size-specific wrapper for the size-generic function. The compiler |
94 | | // will inline the size generic function in here, the advantage is that the size |
95 | | // will be constant allowing for loop unrolling and other constant propagated |
96 | | // goodness. |
97 | | #define CFL_SUBSAMPLE(arch, sub, bd, width, height) \ |
98 | | void cfl_subsample_##bd##_##sub##_##width##x##height##_##arch( \ |
99 | | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3); \ |
100 | | void cfl_subsample_##bd##_##sub##_##width##x##height##_##arch( \ |
101 | 2.59M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ |
102 | 2.59M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ |
103 | 2.59M | output_q3, width, height); \ |
104 | 2.59M | } Unexecuted instantiation: cfl_subsample_lbd_420_4x4_c Unexecuted instantiation: cfl_subsample_lbd_420_8x8_c Unexecuted instantiation: cfl_subsample_lbd_420_16x16_c Unexecuted instantiation: cfl_subsample_lbd_420_32x32_c Unexecuted instantiation: cfl_subsample_lbd_420_4x8_c Unexecuted instantiation: cfl_subsample_lbd_420_8x4_c Unexecuted instantiation: cfl_subsample_lbd_420_8x16_c Unexecuted instantiation: cfl_subsample_lbd_420_16x8_c Unexecuted instantiation: cfl_subsample_lbd_420_16x32_c Unexecuted instantiation: cfl_subsample_lbd_420_32x16_c Unexecuted instantiation: cfl_subsample_lbd_420_4x16_c Unexecuted instantiation: cfl_subsample_lbd_420_16x4_c Unexecuted instantiation: cfl_subsample_lbd_420_8x32_c Unexecuted instantiation: cfl_subsample_lbd_420_32x8_c Unexecuted instantiation: cfl_subsample_lbd_422_4x4_c Unexecuted instantiation: cfl_subsample_lbd_422_8x8_c Unexecuted instantiation: cfl_subsample_lbd_422_16x16_c Unexecuted instantiation: cfl_subsample_lbd_422_32x32_c Unexecuted instantiation: cfl_subsample_lbd_422_4x8_c Unexecuted instantiation: cfl_subsample_lbd_422_8x4_c Unexecuted instantiation: cfl_subsample_lbd_422_8x16_c Unexecuted instantiation: cfl_subsample_lbd_422_16x8_c Unexecuted instantiation: cfl_subsample_lbd_422_16x32_c Unexecuted instantiation: cfl_subsample_lbd_422_32x16_c Unexecuted instantiation: cfl_subsample_lbd_422_4x16_c Unexecuted instantiation: cfl_subsample_lbd_422_16x4_c Unexecuted instantiation: cfl_subsample_lbd_422_8x32_c Unexecuted instantiation: cfl_subsample_lbd_422_32x8_c Unexecuted instantiation: cfl_subsample_lbd_444_4x4_c Unexecuted instantiation: cfl_subsample_lbd_444_8x8_c Unexecuted instantiation: cfl_subsample_lbd_444_16x16_c Unexecuted instantiation: cfl_subsample_lbd_444_32x32_c Unexecuted instantiation: cfl_subsample_lbd_444_4x8_c Unexecuted instantiation: cfl_subsample_lbd_444_8x4_c Unexecuted instantiation: cfl_subsample_lbd_444_8x16_c Unexecuted instantiation: cfl_subsample_lbd_444_16x8_c Unexecuted instantiation: cfl_subsample_lbd_444_16x32_c Unexecuted instantiation: cfl_subsample_lbd_444_32x16_c Unexecuted instantiation: cfl_subsample_lbd_444_4x16_c Unexecuted instantiation: cfl_subsample_lbd_444_16x4_c Unexecuted instantiation: cfl_subsample_lbd_444_8x32_c Unexecuted instantiation: cfl_subsample_lbd_444_32x8_c Unexecuted instantiation: cfl_subsample_hbd_420_4x4_c Unexecuted instantiation: cfl_subsample_hbd_420_8x8_c Unexecuted instantiation: cfl_subsample_hbd_420_16x16_c Unexecuted instantiation: cfl_subsample_hbd_420_32x32_c Unexecuted instantiation: cfl_subsample_hbd_420_4x8_c Unexecuted instantiation: cfl_subsample_hbd_420_8x4_c Unexecuted instantiation: cfl_subsample_hbd_420_8x16_c Unexecuted instantiation: cfl_subsample_hbd_420_16x8_c Unexecuted instantiation: cfl_subsample_hbd_420_16x32_c Unexecuted instantiation: cfl_subsample_hbd_420_32x16_c Unexecuted instantiation: cfl_subsample_hbd_420_4x16_c Unexecuted instantiation: cfl_subsample_hbd_420_16x4_c Unexecuted instantiation: cfl_subsample_hbd_420_8x32_c Unexecuted instantiation: cfl_subsample_hbd_420_32x8_c Unexecuted instantiation: cfl_subsample_hbd_422_4x4_c Unexecuted instantiation: cfl_subsample_hbd_422_8x8_c Unexecuted instantiation: cfl_subsample_hbd_422_16x16_c Unexecuted instantiation: cfl_subsample_hbd_422_32x32_c Unexecuted instantiation: cfl_subsample_hbd_422_4x8_c Unexecuted instantiation: cfl_subsample_hbd_422_8x4_c Unexecuted instantiation: cfl_subsample_hbd_422_8x16_c Unexecuted instantiation: cfl_subsample_hbd_422_16x8_c Unexecuted instantiation: cfl_subsample_hbd_422_16x32_c Unexecuted instantiation: cfl_subsample_hbd_422_32x16_c Unexecuted instantiation: cfl_subsample_hbd_422_4x16_c Unexecuted instantiation: cfl_subsample_hbd_422_16x4_c Unexecuted instantiation: cfl_subsample_hbd_422_8x32_c Unexecuted instantiation: cfl_subsample_hbd_422_32x8_c Unexecuted instantiation: cfl_subsample_hbd_444_4x4_c Unexecuted instantiation: cfl_subsample_hbd_444_8x8_c Unexecuted instantiation: cfl_subsample_hbd_444_16x16_c Unexecuted instantiation: cfl_subsample_hbd_444_32x32_c Unexecuted instantiation: cfl_subsample_hbd_444_4x8_c Unexecuted instantiation: cfl_subsample_hbd_444_8x4_c Unexecuted instantiation: cfl_subsample_hbd_444_8x16_c Unexecuted instantiation: cfl_subsample_hbd_444_16x8_c Unexecuted instantiation: cfl_subsample_hbd_444_16x32_c Unexecuted instantiation: cfl_subsample_hbd_444_32x16_c Unexecuted instantiation: cfl_subsample_hbd_444_4x16_c Unexecuted instantiation: cfl_subsample_hbd_444_16x4_c Unexecuted instantiation: cfl_subsample_hbd_444_8x32_c Unexecuted instantiation: cfl_subsample_hbd_444_32x8_c cfl_subsample_lbd_420_4x4_ssse3 Line | Count | Source | 101 | 136k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 136k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 136k | output_q3, width, height); \ | 104 | 136k | } |
cfl_subsample_lbd_420_8x8_ssse3 Line | Count | Source | 101 | 63.2k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 63.2k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 63.2k | output_q3, width, height); \ | 104 | 63.2k | } |
cfl_subsample_lbd_420_16x16_ssse3 Line | Count | Source | 101 | 34.8k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 34.8k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 34.8k | output_q3, width, height); \ | 104 | 34.8k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x32_ssse3 cfl_subsample_lbd_420_4x8_ssse3 Line | Count | Source | 101 | 74.6k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 74.6k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 74.6k | output_q3, width, height); \ | 104 | 74.6k | } |
cfl_subsample_lbd_420_8x4_ssse3 Line | Count | Source | 101 | 97.8k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 97.8k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 97.8k | output_q3, width, height); \ | 104 | 97.8k | } |
cfl_subsample_lbd_420_8x16_ssse3 Line | Count | Source | 101 | 19.7k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 19.7k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 19.7k | output_q3, width, height); \ | 104 | 19.7k | } |
cfl_subsample_lbd_420_16x8_ssse3 Line | Count | Source | 101 | 31.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 31.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 31.9k | output_q3, width, height); \ | 104 | 31.9k | } |
cfl_subsample_lbd_420_16x32_ssse3 Line | Count | Source | 101 | 6.68k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 6.68k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 6.68k | output_q3, width, height); \ | 104 | 6.68k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x16_ssse3 cfl_subsample_lbd_420_4x16_ssse3 Line | Count | Source | 101 | 80.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 80.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 80.9k | output_q3, width, height); \ | 104 | 80.9k | } |
cfl_subsample_lbd_420_16x4_ssse3 Line | Count | Source | 101 | 142k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 142k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 142k | output_q3, width, height); \ | 104 | 142k | } |
cfl_subsample_lbd_420_8x32_ssse3 Line | Count | Source | 101 | 10.7k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 10.7k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 10.7k | output_q3, width, height); \ | 104 | 10.7k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x8_ssse3 cfl_subsample_lbd_422_4x4_ssse3 Line | Count | Source | 101 | 225 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 225 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 225 | output_q3, width, height); \ | 104 | 225 | } |
cfl_subsample_lbd_422_8x8_ssse3 Line | Count | Source | 101 | 3.65k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 3.65k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 3.65k | output_q3, width, height); \ | 104 | 3.65k | } |
cfl_subsample_lbd_422_16x16_ssse3 Line | Count | Source | 101 | 542 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 542 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 542 | output_q3, width, height); \ | 104 | 542 | } |
Unexecuted instantiation: cfl_subsample_lbd_422_32x32_ssse3 Unexecuted instantiation: cfl_subsample_lbd_422_4x8_ssse3 cfl_subsample_lbd_422_8x4_ssse3 Line | Count | Source | 101 | 215 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 215 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 215 | output_q3, width, height); \ | 104 | 215 | } |
Unexecuted instantiation: cfl_subsample_lbd_422_8x16_ssse3 cfl_subsample_lbd_422_16x8_ssse3 Line | Count | Source | 101 | 284 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 284 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 284 | output_q3, width, height); \ | 104 | 284 | } |
Unexecuted instantiation: cfl_subsample_lbd_422_16x32_ssse3 Unexecuted instantiation: cfl_subsample_lbd_422_32x16_ssse3 Unexecuted instantiation: cfl_subsample_lbd_422_4x16_ssse3 cfl_subsample_lbd_422_16x4_ssse3 Line | Count | Source | 101 | 414 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 414 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 414 | output_q3, width, height); \ | 104 | 414 | } |
Unexecuted instantiation: cfl_subsample_lbd_422_8x32_ssse3 Unexecuted instantiation: cfl_subsample_lbd_422_32x8_ssse3 cfl_subsample_lbd_444_4x4_ssse3 Line | Count | Source | 101 | 41.8k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 41.8k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 41.8k | output_q3, width, height); \ | 104 | 41.8k | } |
cfl_subsample_lbd_444_8x8_ssse3 Line | Count | Source | 101 | 57.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 57.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 57.5k | output_q3, width, height); \ | 104 | 57.5k | } |
cfl_subsample_lbd_444_16x16_ssse3 Line | Count | Source | 101 | 26.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 26.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 26.5k | output_q3, width, height); \ | 104 | 26.5k | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x32_ssse3 cfl_subsample_lbd_444_4x8_ssse3 Line | Count | Source | 101 | 9.66k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 9.66k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 9.66k | output_q3, width, height); \ | 104 | 9.66k | } |
cfl_subsample_lbd_444_8x4_ssse3 Line | Count | Source | 101 | 17.7k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 17.7k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 17.7k | output_q3, width, height); \ | 104 | 17.7k | } |
cfl_subsample_lbd_444_8x16_ssse3 Line | Count | Source | 101 | 16.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 16.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 16.1k | output_q3, width, height); \ | 104 | 16.1k | } |
cfl_subsample_lbd_444_16x8_ssse3 Line | Count | Source | 101 | 25.4k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 25.4k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 25.4k | output_q3, width, height); \ | 104 | 25.4k | } |
cfl_subsample_lbd_444_16x32_ssse3 Line | Count | Source | 101 | 14.0k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14.0k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14.0k | output_q3, width, height); \ | 104 | 14.0k | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x16_ssse3 cfl_subsample_lbd_444_4x16_ssse3 Line | Count | Source | 101 | 141k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 141k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 141k | output_q3, width, height); \ | 104 | 141k | } |
cfl_subsample_lbd_444_16x4_ssse3 Line | Count | Source | 101 | 48.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 48.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 48.1k | output_q3, width, height); \ | 104 | 48.1k | } |
cfl_subsample_lbd_444_8x32_ssse3 Line | Count | Source | 101 | 14.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14.5k | output_q3, width, height); \ | 104 | 14.5k | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x8_ssse3 cfl_subsample_hbd_420_4x4_ssse3 Line | Count | Source | 101 | 166k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 166k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 166k | output_q3, width, height); \ | 104 | 166k | } |
cfl_subsample_hbd_420_8x8_ssse3 Line | Count | Source | 101 | 84.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 84.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 84.1k | output_q3, width, height); \ | 104 | 84.1k | } |
cfl_subsample_hbd_420_16x16_ssse3 Line | Count | Source | 101 | 35.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 35.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 35.5k | output_q3, width, height); \ | 104 | 35.5k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x32_ssse3 cfl_subsample_hbd_420_4x8_ssse3 Line | Count | Source | 101 | 107k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 107k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 107k | output_q3, width, height); \ | 104 | 107k | } |
cfl_subsample_hbd_420_8x4_ssse3 Line | Count | Source | 101 | 138k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 138k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 138k | output_q3, width, height); \ | 104 | 138k | } |
cfl_subsample_hbd_420_8x16_ssse3 Line | Count | Source | 101 | 14.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14.1k | output_q3, width, height); \ | 104 | 14.1k | } |
cfl_subsample_hbd_420_16x8_ssse3 Line | Count | Source | 101 | 33.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 33.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 33.1k | output_q3, width, height); \ | 104 | 33.1k | } |
cfl_subsample_hbd_420_16x32_ssse3 Line | Count | Source | 101 | 4.95k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 4.95k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 4.95k | output_q3, width, height); \ | 104 | 4.95k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x16_ssse3 cfl_subsample_hbd_420_4x16_ssse3 Line | Count | Source | 101 | 131k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 131k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 131k | output_q3, width, height); \ | 104 | 131k | } |
cfl_subsample_hbd_420_16x4_ssse3 Line | Count | Source | 101 | 163k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 163k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 163k | output_q3, width, height); \ | 104 | 163k | } |
cfl_subsample_hbd_420_8x32_ssse3 Line | Count | Source | 101 | 11.0k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 11.0k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 11.0k | output_q3, width, height); \ | 104 | 11.0k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x8_ssse3 cfl_subsample_hbd_422_4x4_ssse3 Line | Count | Source | 101 | 1.85k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 1.85k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 1.85k | output_q3, width, height); \ | 104 | 1.85k | } |
cfl_subsample_hbd_422_8x8_ssse3 Line | Count | Source | 101 | 101 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 101 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 101 | output_q3, width, height); \ | 104 | 101 | } |
cfl_subsample_hbd_422_16x16_ssse3 Line | Count | Source | 101 | 31 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 31 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 31 | output_q3, width, height); \ | 104 | 31 | } |
Unexecuted instantiation: cfl_subsample_hbd_422_32x32_ssse3 Unexecuted instantiation: cfl_subsample_hbd_422_4x8_ssse3 cfl_subsample_hbd_422_8x4_ssse3 Line | Count | Source | 101 | 635 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 635 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 635 | output_q3, width, height); \ | 104 | 635 | } |
Unexecuted instantiation: cfl_subsample_hbd_422_8x16_ssse3 cfl_subsample_hbd_422_16x8_ssse3 Line | Count | Source | 101 | 244 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 244 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 244 | output_q3, width, height); \ | 104 | 244 | } |
Unexecuted instantiation: cfl_subsample_hbd_422_16x32_ssse3 Unexecuted instantiation: cfl_subsample_hbd_422_32x16_ssse3 Unexecuted instantiation: cfl_subsample_hbd_422_4x16_ssse3 cfl_subsample_hbd_422_16x4_ssse3 Line | Count | Source | 101 | 45 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 45 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 45 | output_q3, width, height); \ | 104 | 45 | } |
Unexecuted instantiation: cfl_subsample_hbd_422_8x32_ssse3 Unexecuted instantiation: cfl_subsample_hbd_422_32x8_ssse3 cfl_subsample_hbd_444_4x4_ssse3 Line | Count | Source | 101 | 80.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 80.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 80.1k | output_q3, width, height); \ | 104 | 80.1k | } |
cfl_subsample_hbd_444_8x8_ssse3 Line | Count | Source | 101 | 102k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 102k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 102k | output_q3, width, height); \ | 104 | 102k | } |
cfl_subsample_hbd_444_16x16_ssse3 Line | Count | Source | 101 | 35.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 35.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 35.9k | output_q3, width, height); \ | 104 | 35.9k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x32_ssse3 cfl_subsample_hbd_444_4x8_ssse3 Line | Count | Source | 101 | 22.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 22.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 22.5k | output_q3, width, height); \ | 104 | 22.5k | } |
cfl_subsample_hbd_444_8x4_ssse3 Line | Count | Source | 101 | 36.3k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 36.3k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 36.3k | output_q3, width, height); \ | 104 | 36.3k | } |
cfl_subsample_hbd_444_8x16_ssse3 Line | Count | Source | 101 | 30.3k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 30.3k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 30.3k | output_q3, width, height); \ | 104 | 30.3k | } |
cfl_subsample_hbd_444_16x8_ssse3 Line | Count | Source | 101 | 38.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 38.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 38.5k | output_q3, width, height); \ | 104 | 38.5k | } |
cfl_subsample_hbd_444_16x32_ssse3 Line | Count | Source | 101 | 6.76k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 6.76k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 6.76k | output_q3, width, height); \ | 104 | 6.76k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x16_ssse3 cfl_subsample_hbd_444_4x16_ssse3 Line | Count | Source | 101 | 22.6k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 22.6k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 22.6k | output_q3, width, height); \ | 104 | 22.6k | } |
cfl_subsample_hbd_444_16x4_ssse3 Line | Count | Source | 101 | 38.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 38.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 38.1k | output_q3, width, height); \ | 104 | 38.1k | } |
cfl_subsample_hbd_444_8x32_ssse3 Line | Count | Source | 101 | 15.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 15.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 15.1k | output_q3, width, height); \ | 104 | 15.1k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x8_ssse3 cfl_subsample_lbd_420_32x32_avx2 Line | Count | Source | 101 | 20.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 20.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 20.5k | output_q3, width, height); \ | 104 | 20.5k | } |
cfl_subsample_lbd_420_32x16_avx2 Line | Count | Source | 101 | 6.46k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 6.46k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 6.46k | output_q3, width, height); \ | 104 | 6.46k | } |
cfl_subsample_lbd_420_32x8_avx2 Line | Count | Source | 101 | 11.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 11.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 11.9k | output_q3, width, height); \ | 104 | 11.9k | } |
cfl_subsample_lbd_422_32x32_avx2 Line | Count | Source | 101 | 445 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 445 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 445 | output_q3, width, height); \ | 104 | 445 | } |
cfl_subsample_lbd_422_32x16_avx2 Line | Count | Source | 101 | 166 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 166 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 166 | output_q3, width, height); \ | 104 | 166 | } |
cfl_subsample_lbd_422_32x8_avx2 Line | Count | Source | 101 | 391 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 391 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 391 | output_q3, width, height); \ | 104 | 391 | } |
cfl_subsample_lbd_444_32x32_avx2 Line | Count | Source | 101 | 14.6k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14.6k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14.6k | output_q3, width, height); \ | 104 | 14.6k | } |
cfl_subsample_lbd_444_32x16_avx2 Line | Count | Source | 101 | 6.04k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 6.04k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 6.04k | output_q3, width, height); \ | 104 | 6.04k | } |
cfl_subsample_lbd_444_32x8_avx2 Line | Count | Source | 101 | 12.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 12.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 12.1k | output_q3, width, height); \ | 104 | 12.1k | } |
cfl_subsample_hbd_420_32x32_avx2 Line | Count | Source | 101 | 17.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 17.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 17.5k | output_q3, width, height); \ | 104 | 17.5k | } |
cfl_subsample_hbd_420_32x16_avx2 Line | Count | Source | 101 | 6.67k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 6.67k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 6.67k | output_q3, width, height); \ | 104 | 6.67k | } |
cfl_subsample_hbd_420_32x8_avx2 Line | Count | Source | 101 | 10.0k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 10.0k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 10.0k | output_q3, width, height); \ | 104 | 10.0k | } |
cfl_subsample_hbd_422_32x32_avx2 Line | Count | Source | 101 | 14 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14 | output_q3, width, height); \ | 104 | 14 | } |
cfl_subsample_hbd_422_32x16_avx2 Line | Count | Source | 101 | 126 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 126 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 126 | output_q3, width, height); \ | 104 | 126 | } |
cfl_subsample_hbd_422_32x8_avx2 Line | Count | Source | 101 | 63 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 63 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 63 | output_q3, width, height); \ | 104 | 63 | } |
cfl_subsample_hbd_444_32x32_avx2 Line | Count | Source | 101 | 14.7k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 14.7k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 14.7k | output_q3, width, height); \ | 104 | 14.7k | } |
cfl_subsample_hbd_444_32x16_avx2 Line | Count | Source | 101 | 10.3k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 10.3k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 10.3k | output_q3, width, height); \ | 104 | 10.3k | } |
cfl_subsample_hbd_444_32x8_avx2 Line | Count | Source | 101 | 17.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 102 | 17.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 103 | 17.5k | output_q3, width, height); \ | 104 | 17.5k | } |
|
105 | | |
106 | | // Declare size-specific wrappers for all valid CfL sizes. |
107 | | #define CFL_SUBSAMPLE_FUNCTIONS(arch, sub, bd) \ |
108 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 4) \ |
109 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 8) \ |
110 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 16) \ |
111 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 32) \ |
112 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 8) \ |
113 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 4) \ |
114 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 16) \ |
115 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 8) \ |
116 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 32) \ |
117 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 16) \ |
118 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 16) \ |
119 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 4) \ |
120 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 32) \ |
121 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 8) \ |
122 | | cfl_subsample_##bd##_fn cfl_get_luma_subsampling_##sub##_##bd##_##arch( \ |
123 | 0 | TX_SIZE tx_size) { \ |
124 | 0 | CFL_SUBSAMPLE_FUNCTION_ARRAY(arch, sub, bd) \ |
125 | 0 | return subfn_##sub[tx_size]; \ |
126 | 0 | } |
127 | | |
128 | | // Declare an architecture-specific array of function pointers for size-specific |
129 | | // wrappers. |
130 | | #define CFL_SUBSAMPLE_FUNCTION_ARRAY(arch, sub, bd) \ |
131 | 0 | static const cfl_subsample_##bd##_fn subfn_##sub[TX_SIZES_ALL] = { \ |
132 | 0 | cfl_subsample_##bd##_##sub##_4x4_##arch, /* 4x4 */ \ |
133 | 0 | cfl_subsample_##bd##_##sub##_8x8_##arch, /* 8x8 */ \ |
134 | 0 | cfl_subsample_##bd##_##sub##_16x16_##arch, /* 16x16 */ \ |
135 | 0 | cfl_subsample_##bd##_##sub##_32x32_##arch, /* 32x32 */ \ |
136 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
137 | 0 | cfl_subsample_##bd##_##sub##_4x8_##arch, /* 4x8 */ \ |
138 | 0 | cfl_subsample_##bd##_##sub##_8x4_##arch, /* 8x4 */ \ |
139 | 0 | cfl_subsample_##bd##_##sub##_8x16_##arch, /* 8x16 */ \ |
140 | 0 | cfl_subsample_##bd##_##sub##_16x8_##arch, /* 16x8 */ \ |
141 | 0 | cfl_subsample_##bd##_##sub##_16x32_##arch, /* 16x32 */ \ |
142 | 0 | cfl_subsample_##bd##_##sub##_32x16_##arch, /* 32x16 */ \ |
143 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
144 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
145 | 0 | cfl_subsample_##bd##_##sub##_4x16_##arch, /* 4x16 */ \ |
146 | 0 | cfl_subsample_##bd##_##sub##_16x4_##arch, /* 16x4 */ \ |
147 | 0 | cfl_subsample_##bd##_##sub##_8x32_##arch, /* 8x32 */ \ |
148 | 0 | cfl_subsample_##bd##_##sub##_32x8_##arch, /* 32x8 */ \ |
149 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
150 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
151 | 0 | }; |
152 | | |
153 | | // The RTCD script does not support passing in an array, so we wrap it in this |
154 | | // function. |
155 | | #if CONFIG_AV1_HIGHBITDEPTH |
156 | | #define CFL_GET_SUBSAMPLE_FUNCTION(arch) \ |
157 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 420, lbd) \ |
158 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 422, lbd) \ |
159 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 444, lbd) \ |
160 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 420, hbd) \ |
161 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 422, hbd) \ |
162 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 444, hbd) |
163 | | #else |
164 | | #define CFL_GET_SUBSAMPLE_FUNCTION(arch) \ |
165 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 420, lbd) \ |
166 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 422, lbd) \ |
167 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 444, lbd) |
168 | | #endif |
169 | | |
170 | | // Declare a size-specific wrapper for the size-generic function. The compiler |
171 | | // will inline the size generic function in here, the advantage is that the size |
172 | | // will be constant allowing for loop unrolling and other constant propagated |
173 | | // goodness. |
174 | | #define CFL_SUB_AVG_X(arch, width, height, round_offset, num_pel_log2) \ |
175 | | void cfl_subtract_average_##width##x##height##_##arch(const uint16_t *src, \ |
176 | | int16_t *dst); \ |
177 | | void cfl_subtract_average_##width##x##height##_##arch(const uint16_t *src, \ |
178 | 1.32M | int16_t *dst) { \ |
179 | 1.32M | subtract_average_##arch(src, dst, width, height, round_offset, \ |
180 | 1.32M | num_pel_log2); \ |
181 | 1.32M | } Unexecuted instantiation: cfl_subtract_average_4x4_c Unexecuted instantiation: cfl_subtract_average_4x8_c Unexecuted instantiation: cfl_subtract_average_4x16_c Unexecuted instantiation: cfl_subtract_average_8x4_c Unexecuted instantiation: cfl_subtract_average_8x8_c Unexecuted instantiation: cfl_subtract_average_8x16_c Unexecuted instantiation: cfl_subtract_average_8x32_c Unexecuted instantiation: cfl_subtract_average_16x4_c Unexecuted instantiation: cfl_subtract_average_16x8_c Unexecuted instantiation: cfl_subtract_average_16x16_c Unexecuted instantiation: cfl_subtract_average_16x32_c Unexecuted instantiation: cfl_subtract_average_32x8_c Unexecuted instantiation: cfl_subtract_average_32x16_c Unexecuted instantiation: cfl_subtract_average_32x32_c cfl_subtract_average_4x4_sse2 Line | Count | Source | 178 | 227k | int16_t *dst) { \ | 179 | 227k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 227k | num_pel_log2); \ | 181 | 227k | } |
cfl_subtract_average_4x8_sse2 Line | Count | Source | 178 | 84.2k | int16_t *dst) { \ | 179 | 84.2k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 84.2k | num_pel_log2); \ | 181 | 84.2k | } |
cfl_subtract_average_4x16_sse2 Line | Count | Source | 178 | 190k | int16_t *dst) { \ | 179 | 190k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 190k | num_pel_log2); \ | 181 | 190k | } |
cfl_subtract_average_8x4_sse2 Line | Count | Source | 178 | 142k | int16_t *dst) { \ | 179 | 142k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 142k | num_pel_log2); \ | 181 | 142k | } |
cfl_subtract_average_8x8_sse2 Line | Count | Source | 178 | 188k | int16_t *dst) { \ | 179 | 188k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 188k | num_pel_log2); \ | 181 | 188k | } |
cfl_subtract_average_8x16_sse2 Line | Count | Source | 178 | 59.8k | int16_t *dst) { \ | 179 | 59.8k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 59.8k | num_pel_log2); \ | 181 | 59.8k | } |
cfl_subtract_average_8x32_sse2 Line | Count | Source | 178 | 33.0k | int16_t *dst) { \ | 179 | 33.0k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 33.0k | num_pel_log2); \ | 181 | 33.0k | } |
Unexecuted instantiation: cfl_subtract_average_16x4_sse2 Unexecuted instantiation: cfl_subtract_average_16x8_sse2 Unexecuted instantiation: cfl_subtract_average_16x16_sse2 Unexecuted instantiation: cfl_subtract_average_16x32_sse2 Unexecuted instantiation: cfl_subtract_average_32x8_sse2 Unexecuted instantiation: cfl_subtract_average_32x16_sse2 Unexecuted instantiation: cfl_subtract_average_32x32_sse2 cfl_subtract_average_16x4_avx2 Line | Count | Source | 178 | 113k | int16_t *dst) { \ | 179 | 113k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 113k | num_pel_log2); \ | 181 | 113k | } |
cfl_subtract_average_16x8_avx2 Line | Count | Source | 178 | 79.8k | int16_t *dst) { \ | 179 | 79.8k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 79.8k | num_pel_log2); \ | 181 | 79.8k | } |
cfl_subtract_average_16x16_avx2 Line | Count | Source | 178 | 106k | int16_t *dst) { \ | 179 | 106k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 106k | num_pel_log2); \ | 181 | 106k | } |
cfl_subtract_average_16x32_avx2 Line | Count | Source | 178 | 22.0k | int16_t *dst) { \ | 179 | 22.0k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 22.0k | num_pel_log2); \ | 181 | 22.0k | } |
cfl_subtract_average_32x8_avx2 Line | Count | Source | 178 | 30.1k | int16_t *dst) { \ | 179 | 30.1k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 30.1k | num_pel_log2); \ | 181 | 30.1k | } |
cfl_subtract_average_32x16_avx2 Line | Count | Source | 178 | 16.6k | int16_t *dst) { \ | 179 | 16.6k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 16.6k | num_pel_log2); \ | 181 | 16.6k | } |
cfl_subtract_average_32x32_avx2 Line | Count | Source | 178 | 29.9k | int16_t *dst) { \ | 179 | 29.9k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 180 | 29.9k | num_pel_log2); \ | 181 | 29.9k | } |
|
182 | | |
183 | | // Declare size-specific wrappers for all valid CfL sizes. |
184 | | #define CFL_SUB_AVG_FN(arch) \ |
185 | | CFL_SUB_AVG_X(arch, 4, 4, 8, 4) \ |
186 | | CFL_SUB_AVG_X(arch, 4, 8, 16, 5) \ |
187 | | CFL_SUB_AVG_X(arch, 4, 16, 32, 6) \ |
188 | | CFL_SUB_AVG_X(arch, 8, 4, 16, 5) \ |
189 | | CFL_SUB_AVG_X(arch, 8, 8, 32, 6) \ |
190 | | CFL_SUB_AVG_X(arch, 8, 16, 64, 7) \ |
191 | | CFL_SUB_AVG_X(arch, 8, 32, 128, 8) \ |
192 | | CFL_SUB_AVG_X(arch, 16, 4, 32, 6) \ |
193 | | CFL_SUB_AVG_X(arch, 16, 8, 64, 7) \ |
194 | | CFL_SUB_AVG_X(arch, 16, 16, 128, 8) \ |
195 | | CFL_SUB_AVG_X(arch, 16, 32, 256, 9) \ |
196 | | CFL_SUB_AVG_X(arch, 32, 8, 128, 8) \ |
197 | | CFL_SUB_AVG_X(arch, 32, 16, 256, 9) \ |
198 | | CFL_SUB_AVG_X(arch, 32, 32, 512, 10) \ |
199 | | cfl_subtract_average_fn cfl_get_subtract_average_fn_##arch( \ |
200 | 0 | TX_SIZE tx_size) { \ |
201 | 0 | static const cfl_subtract_average_fn sub_avg[TX_SIZES_ALL] = { \ |
202 | 0 | cfl_subtract_average_4x4_##arch, /* 4x4 */ \ |
203 | 0 | cfl_subtract_average_8x8_##arch, /* 8x8 */ \ |
204 | 0 | cfl_subtract_average_16x16_##arch, /* 16x16 */ \ |
205 | 0 | cfl_subtract_average_32x32_##arch, /* 32x32 */ \ |
206 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
207 | 0 | cfl_subtract_average_4x8_##arch, /* 4x8 */ \ |
208 | 0 | cfl_subtract_average_8x4_##arch, /* 8x4 */ \ |
209 | 0 | cfl_subtract_average_8x16_##arch, /* 8x16 */ \ |
210 | 0 | cfl_subtract_average_16x8_##arch, /* 16x8 */ \ |
211 | 0 | cfl_subtract_average_16x32_##arch, /* 16x32 */ \ |
212 | 0 | cfl_subtract_average_32x16_##arch, /* 32x16 */ \ |
213 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
214 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
215 | 0 | cfl_subtract_average_4x16_##arch, /* 4x16 (invalid CFL size) */ \ |
216 | 0 | cfl_subtract_average_16x4_##arch, /* 16x4 (invalid CFL size) */ \ |
217 | 0 | cfl_subtract_average_8x32_##arch, /* 8x32 (invalid CFL size) */ \ |
218 | 0 | cfl_subtract_average_32x8_##arch, /* 32x8 (invalid CFL size) */ \ |
219 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
220 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
221 | 0 | }; \ |
222 | 0 | /* Modulo TX_SIZES_ALL to ensure that an attacker won't be able to */ \ |
223 | 0 | /* index the function pointer array out of bounds. */ \ |
224 | 0 | return sub_avg[tx_size % TX_SIZES_ALL]; \ |
225 | 0 | } |
226 | | |
227 | | #define CFL_PREDICT_lbd(arch, width, height) \ |
228 | | void cfl_predict_lbd_##width##x##height##_##arch( \ |
229 | | const int16_t *pred_buf_q3, uint8_t *dst, int dst_stride, int alpha_q3); \ |
230 | | void cfl_predict_lbd_##width##x##height##_##arch( \ |
231 | | const int16_t *pred_buf_q3, uint8_t *dst, int dst_stride, \ |
232 | 1.35M | int alpha_q3) { \ |
233 | 1.35M | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ |
234 | 1.35M | height); \ |
235 | 1.35M | } Unexecuted instantiation: cfl_predict_lbd_4x4_c Unexecuted instantiation: cfl_predict_lbd_4x8_c Unexecuted instantiation: cfl_predict_lbd_4x16_c Unexecuted instantiation: cfl_predict_lbd_8x4_c Unexecuted instantiation: cfl_predict_lbd_8x8_c Unexecuted instantiation: cfl_predict_lbd_8x16_c Unexecuted instantiation: cfl_predict_lbd_8x32_c Unexecuted instantiation: cfl_predict_lbd_16x4_c Unexecuted instantiation: cfl_predict_lbd_16x8_c Unexecuted instantiation: cfl_predict_lbd_16x16_c Unexecuted instantiation: cfl_predict_lbd_16x32_c Unexecuted instantiation: cfl_predict_lbd_32x8_c Unexecuted instantiation: cfl_predict_lbd_32x16_c Unexecuted instantiation: cfl_predict_lbd_32x32_c cfl_predict_lbd_4x4_ssse3 Line | Count | Source | 232 | 215k | int alpha_q3) { \ | 233 | 215k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 215k | height); \ | 235 | 215k | } |
cfl_predict_lbd_4x8_ssse3 Line | Count | Source | 232 | 79.0k | int alpha_q3) { \ | 233 | 79.0k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 79.0k | height); \ | 235 | 79.0k | } |
cfl_predict_lbd_4x16_ssse3 Line | Count | Source | 232 | 305k | int alpha_q3) { \ | 233 | 305k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 305k | height); \ | 235 | 305k | } |
cfl_predict_lbd_8x4_ssse3 Line | Count | Source | 232 | 129k | int alpha_q3) { \ | 233 | 129k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 129k | height); \ | 235 | 129k | } |
cfl_predict_lbd_8x8_ssse3 Line | Count | Source | 232 | 170k | int alpha_q3) { \ | 233 | 170k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 170k | height); \ | 235 | 170k | } |
cfl_predict_lbd_8x16_ssse3 Line | Count | Source | 232 | 48.0k | int alpha_q3) { \ | 233 | 48.0k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 48.0k | height); \ | 235 | 48.0k | } |
cfl_predict_lbd_8x32_ssse3 Line | Count | Source | 232 | 30.5k | int alpha_q3) { \ | 233 | 30.5k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 30.5k | height); \ | 235 | 30.5k | } |
cfl_predict_lbd_16x4_ssse3 Line | Count | Source | 232 | 121k | int alpha_q3) { \ | 233 | 121k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 121k | height); \ | 235 | 121k | } |
cfl_predict_lbd_16x8_ssse3 Line | Count | Source | 232 | 66.0k | int alpha_q3) { \ | 233 | 66.0k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 66.0k | height); \ | 235 | 66.0k | } |
cfl_predict_lbd_16x16_ssse3 Line | Count | Source | 232 | 96.5k | int alpha_q3) { \ | 233 | 96.5k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 96.5k | height); \ | 235 | 96.5k | } |
cfl_predict_lbd_16x32_ssse3 Line | Count | Source | 232 | 29.1k | int alpha_q3) { \ | 233 | 29.1k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 29.1k | height); \ | 235 | 29.1k | } |
Unexecuted instantiation: cfl_predict_lbd_32x8_ssse3 Unexecuted instantiation: cfl_predict_lbd_32x16_ssse3 Unexecuted instantiation: cfl_predict_lbd_32x32_ssse3 cfl_predict_lbd_32x8_avx2 Line | Count | Source | 232 | 24.3k | int alpha_q3) { \ | 233 | 24.3k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 24.3k | height); \ | 235 | 24.3k | } |
cfl_predict_lbd_32x16_avx2 Line | Count | Source | 232 | 12.2k | int alpha_q3) { \ | 233 | 12.2k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 12.2k | height); \ | 235 | 12.2k | } |
cfl_predict_lbd_32x32_avx2 Line | Count | Source | 232 | 29.6k | int alpha_q3) { \ | 233 | 29.6k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 234 | 29.6k | height); \ | 235 | 29.6k | } |
|
236 | | |
237 | | #if CONFIG_AV1_HIGHBITDEPTH |
238 | | #define CFL_PREDICT_hbd(arch, width, height) \ |
239 | | void cfl_predict_hbd_##width##x##height##_##arch( \ |
240 | | const int16_t *pred_buf_q3, uint16_t *dst, int dst_stride, int alpha_q3, \ |
241 | | int bd); \ |
242 | | void cfl_predict_hbd_##width##x##height##_##arch( \ |
243 | | const int16_t *pred_buf_q3, uint16_t *dst, int dst_stride, int alpha_q3, \ |
244 | 1.29M | int bd) { \ |
245 | 1.29M | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ |
246 | 1.29M | height); \ |
247 | 1.29M | } Unexecuted instantiation: cfl_predict_hbd_4x4_c Unexecuted instantiation: cfl_predict_hbd_4x8_c Unexecuted instantiation: cfl_predict_hbd_4x16_c Unexecuted instantiation: cfl_predict_hbd_8x4_c Unexecuted instantiation: cfl_predict_hbd_8x8_c Unexecuted instantiation: cfl_predict_hbd_8x16_c Unexecuted instantiation: cfl_predict_hbd_8x32_c Unexecuted instantiation: cfl_predict_hbd_16x4_c Unexecuted instantiation: cfl_predict_hbd_16x8_c Unexecuted instantiation: cfl_predict_hbd_16x16_c Unexecuted instantiation: cfl_predict_hbd_16x32_c Unexecuted instantiation: cfl_predict_hbd_32x8_c Unexecuted instantiation: cfl_predict_hbd_32x16_c Unexecuted instantiation: cfl_predict_hbd_32x32_c cfl_predict_hbd_4x4_ssse3 Line | Count | Source | 244 | 240k | int bd) { \ | 245 | 240k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 240k | height); \ | 247 | 240k | } |
cfl_predict_hbd_4x8_ssse3 Line | Count | Source | 244 | 89.4k | int bd) { \ | 245 | 89.4k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 89.4k | height); \ | 247 | 89.4k | } |
cfl_predict_hbd_4x16_ssse3 Line | Count | Source | 244 | 74.2k | int bd) { \ | 245 | 74.2k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 74.2k | height); \ | 247 | 74.2k | } |
cfl_predict_hbd_8x4_ssse3 Line | Count | Source | 244 | 156k | int bd) { \ | 245 | 156k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 156k | height); \ | 247 | 156k | } |
cfl_predict_hbd_8x8_ssse3 Line | Count | Source | 244 | 206k | int bd) { \ | 245 | 206k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 206k | height); \ | 247 | 206k | } |
cfl_predict_hbd_8x16_ssse3 Line | Count | Source | 244 | 71.7k | int bd) { \ | 245 | 71.7k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 71.7k | height); \ | 247 | 71.7k | } |
cfl_predict_hbd_8x32_ssse3 Line | Count | Source | 244 | 35.4k | int bd) { \ | 245 | 35.4k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 35.4k | height); \ | 247 | 35.4k | } |
Unexecuted instantiation: cfl_predict_hbd_16x4_ssse3 Unexecuted instantiation: cfl_predict_hbd_16x8_ssse3 Unexecuted instantiation: cfl_predict_hbd_16x16_ssse3 Unexecuted instantiation: cfl_predict_hbd_16x32_ssse3 Unexecuted instantiation: cfl_predict_hbd_32x8_ssse3 Unexecuted instantiation: cfl_predict_hbd_32x16_ssse3 Unexecuted instantiation: cfl_predict_hbd_32x32_ssse3 cfl_predict_hbd_16x4_avx2 Line | Count | Source | 244 | 106k | int bd) { \ | 245 | 106k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 106k | height); \ | 247 | 106k | } |
cfl_predict_hbd_16x8_avx2 Line | Count | Source | 244 | 93.6k | int bd) { \ | 245 | 93.6k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 93.6k | height); \ | 247 | 93.6k | } |
cfl_predict_hbd_16x16_avx2 Line | Count | Source | 244 | 116k | int bd) { \ | 245 | 116k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 116k | height); \ | 247 | 116k | } |
cfl_predict_hbd_16x32_avx2 Line | Count | Source | 244 | 14.9k | int bd) { \ | 245 | 14.9k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 14.9k | height); \ | 247 | 14.9k | } |
cfl_predict_hbd_32x8_avx2 Line | Count | Source | 244 | 35.9k | int bd) { \ | 245 | 35.9k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 35.9k | height); \ | 247 | 35.9k | } |
cfl_predict_hbd_32x16_avx2 Line | Count | Source | 244 | 21.0k | int bd) { \ | 245 | 21.0k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 21.0k | height); \ | 247 | 21.0k | } |
cfl_predict_hbd_32x32_avx2 Line | Count | Source | 244 | 30.1k | int bd) { \ | 245 | 30.1k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 246 | 30.1k | height); \ | 247 | 30.1k | } |
|
248 | | #endif |
249 | | |
250 | | // This wrapper exists because clang format does not like calling macros with |
251 | | // lowercase letters. |
252 | | #define CFL_PREDICT_X(arch, width, height, bd) \ |
253 | | CFL_PREDICT_##bd(arch, width, height) |
254 | | |
255 | | #define CFL_PREDICT_FN(arch, bd) \ |
256 | | CFL_PREDICT_X(arch, 4, 4, bd) \ |
257 | | CFL_PREDICT_X(arch, 4, 8, bd) \ |
258 | | CFL_PREDICT_X(arch, 4, 16, bd) \ |
259 | | CFL_PREDICT_X(arch, 8, 4, bd) \ |
260 | | CFL_PREDICT_X(arch, 8, 8, bd) \ |
261 | | CFL_PREDICT_X(arch, 8, 16, bd) \ |
262 | | CFL_PREDICT_X(arch, 8, 32, bd) \ |
263 | | CFL_PREDICT_X(arch, 16, 4, bd) \ |
264 | | CFL_PREDICT_X(arch, 16, 8, bd) \ |
265 | | CFL_PREDICT_X(arch, 16, 16, bd) \ |
266 | | CFL_PREDICT_X(arch, 16, 32, bd) \ |
267 | | CFL_PREDICT_X(arch, 32, 8, bd) \ |
268 | | CFL_PREDICT_X(arch, 32, 16, bd) \ |
269 | | CFL_PREDICT_X(arch, 32, 32, bd) \ |
270 | 0 | cfl_predict_##bd##_fn cfl_get_predict_##bd##_fn_##arch(TX_SIZE tx_size) { \ |
271 | 0 | static const cfl_predict_##bd##_fn pred[TX_SIZES_ALL] = { \ |
272 | 0 | cfl_predict_##bd##_4x4_##arch, /* 4x4 */ \ |
273 | 0 | cfl_predict_##bd##_8x8_##arch, /* 8x8 */ \ |
274 | 0 | cfl_predict_##bd##_16x16_##arch, /* 16x16 */ \ |
275 | 0 | cfl_predict_##bd##_32x32_##arch, /* 32x32 */ \ |
276 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
277 | 0 | cfl_predict_##bd##_4x8_##arch, /* 4x8 */ \ |
278 | 0 | cfl_predict_##bd##_8x4_##arch, /* 8x4 */ \ |
279 | 0 | cfl_predict_##bd##_8x16_##arch, /* 8x16 */ \ |
280 | 0 | cfl_predict_##bd##_16x8_##arch, /* 16x8 */ \ |
281 | 0 | cfl_predict_##bd##_16x32_##arch, /* 16x32 */ \ |
282 | 0 | cfl_predict_##bd##_32x16_##arch, /* 32x16 */ \ |
283 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
284 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
285 | 0 | cfl_predict_##bd##_4x16_##arch, /* 4x16 */ \ |
286 | 0 | cfl_predict_##bd##_16x4_##arch, /* 16x4 */ \ |
287 | 0 | cfl_predict_##bd##_8x32_##arch, /* 8x32 */ \ |
288 | 0 | cfl_predict_##bd##_32x8_##arch, /* 32x8 */ \ |
289 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
290 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
291 | 0 | }; \ |
292 | 0 | /* Modulo TX_SIZES_ALL to ensure that an attacker won't be able to */ \ |
293 | 0 | /* index the function pointer array out of bounds. */ \ |
294 | 0 | return pred[tx_size % TX_SIZES_ALL]; \ |
295 | 0 | } |
296 | | |
297 | | #endif // AOM_AV1_COMMON_CFL_H_ |