/src/aom/av1/common/cfl.h
Line | Count | Source (jump to first uncovered line) |
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 | 34.1M | static INLINE CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { |
20 | 34.1M | const MB_MODE_INFO *mbmi = xd->mi[0]; |
21 | 34.1M | const BLOCK_SIZE bsize = mbmi->bsize; |
22 | 34.1M | assert(bsize < BLOCK_SIZES_ALL); |
23 | 34.1M | if (xd->lossless[mbmi->segment_id]) { |
24 | | // In lossless, CfL is available when the partition size is equal to the |
25 | | // transform size. |
26 | 72.9k | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; |
27 | 72.9k | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; |
28 | 72.9k | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); |
29 | 72.9k | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); |
30 | 72.9k | } |
31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 |
32 | 34.0M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && |
33 | 34.0M | block_size_high[bsize] <= 32); |
34 | 34.1M | } Unexecuted instantiation: decodeframe.c:is_cfl_allowed decodemv.c:is_cfl_allowed Line | Count | Source | 19 | 29.6M | static INLINE CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { | 20 | 29.6M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 21 | 29.6M | const BLOCK_SIZE bsize = mbmi->bsize; | 22 | 29.6M | assert(bsize < BLOCK_SIZES_ALL); | 23 | 29.6M | if (xd->lossless[mbmi->segment_id]) { | 24 | | // In lossless, CfL is available when the partition size is equal to the | 25 | | // transform size. | 26 | 69.0k | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; | 27 | 69.0k | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; | 28 | 69.0k | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); | 29 | 69.0k | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); | 30 | 69.0k | } | 31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 | 32 | 29.6M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && | 33 | 29.6M | block_size_high[bsize] <= 32); | 34 | 29.6M | } |
Line | Count | Source | 19 | 4.42M | static INLINE CFL_ALLOWED_TYPE is_cfl_allowed(const MACROBLOCKD *xd) { | 20 | 4.42M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 21 | 4.42M | const BLOCK_SIZE bsize = mbmi->bsize; | 22 | 4.42M | assert(bsize < BLOCK_SIZES_ALL); | 23 | 4.42M | if (xd->lossless[mbmi->segment_id]) { | 24 | | // In lossless, CfL is available when the partition size is equal to the | 25 | | // transform size. | 26 | 3.97k | const int ssx = xd->plane[AOM_PLANE_U].subsampling_x; | 27 | 3.97k | const int ssy = xd->plane[AOM_PLANE_U].subsampling_y; | 28 | 3.97k | const int plane_bsize = get_plane_block_size(bsize, ssx, ssy); | 29 | 3.97k | return (CFL_ALLOWED_TYPE)(plane_bsize == BLOCK_4X4); | 30 | 3.97k | } | 31 | | // Spec: CfL is available to luma partitions lesser than or equal to 32x32 | 32 | 4.41M | return (CFL_ALLOWED_TYPE)(block_size_wide[bsize] <= 32 && | 33 | 4.41M | block_size_high[bsize] <= 32); | 34 | 4.42M | } |
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 | 178M | const MACROBLOCKD *xd) { |
40 | 178M | const MB_MODE_INFO *mbmi = xd->mi[0]; |
41 | | |
42 | 178M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; |
43 | | |
44 | 138M | 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 | 30.1M | return CFL_ALLOWED; |
51 | 30.1M | } |
52 | | |
53 | | // If this block has chroma information, we know whether we're |
54 | | // actually going to perform a CfL prediction |
55 | 108M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && |
56 | 108M | mbmi->uv_mode == UV_CFL_PRED); |
57 | 138M | } decodeframe.c:store_cfl_required Line | Count | Source | 39 | 96.6M | const MACROBLOCKD *xd) { | 40 | 96.6M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 41 | | | 42 | 96.6M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; | 43 | | | 44 | 73.4M | 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 | 15.2M | return CFL_ALLOWED; | 51 | 15.2M | } | 52 | | | 53 | | // If this block has chroma information, we know whether we're | 54 | | // actually going to perform a CfL prediction | 55 | 58.1M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && | 56 | 58.1M | mbmi->uv_mode == UV_CFL_PRED); | 57 | 73.4M | } |
decodemv.c:store_cfl_required Line | Count | Source | 39 | 81.9M | const MACROBLOCKD *xd) { | 40 | 81.9M | const MB_MODE_INFO *mbmi = xd->mi[0]; | 41 | | | 42 | 81.9M | if (cm->seq_params->monochrome) return CFL_DISALLOWED; | 43 | | | 44 | 64.8M | 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 | 14.9M | return CFL_ALLOWED; | 51 | 14.9M | } | 52 | | | 53 | | // If this block has chroma information, we know whether we're | 54 | | // actually going to perform a CfL prediction | 55 | 49.8M | return (CFL_ALLOWED_TYPE)(!is_inter_block(mbmi) && | 56 | 49.8M | mbmi->uv_mode == UV_CFL_PRED); | 57 | 64.8M | } |
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 | 4.42M | static INLINE CFL_PRED_TYPE get_cfl_pred_type(PLANE_TYPE plane) { |
65 | 4.42M | assert(plane > 0); |
66 | 0 | return (CFL_PRED_TYPE)(plane - 1); |
67 | 4.42M | } 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 | 4.42M | static INLINE CFL_PRED_TYPE get_cfl_pred_type(PLANE_TYPE plane) { | 65 | 4.42M | assert(plane > 0); | 66 | 0 | return (CFL_PRED_TYPE)(plane - 1); | 67 | 4.42M | } |
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 | | void cfl_predict_block(MACROBLOCKD *const xd, uint8_t *dst, int dst_stride, |
70 | | TX_SIZE tx_size, int plane); |
71 | | |
72 | | void cfl_store_block(MACROBLOCKD *const xd, BLOCK_SIZE bsize, TX_SIZE tx_size); |
73 | | |
74 | | void cfl_store_tx(MACROBLOCKD *const xd, int row, int col, TX_SIZE tx_size, |
75 | | BLOCK_SIZE bsize); |
76 | | |
77 | | void cfl_store_dc_pred(MACROBLOCKD *const xd, const uint8_t *input, |
78 | | CFL_PRED_TYPE pred_plane, int width); |
79 | | |
80 | | void cfl_load_dc_pred(MACROBLOCKD *const xd, uint8_t *dst, int dst_stride, |
81 | | TX_SIZE tx_size, CFL_PRED_TYPE pred_plane); |
82 | | |
83 | | // Allows the CFL_SUBSAMPLE function to switch types depending on the bitdepth. |
84 | | #define CFL_lbd_TYPE uint8_t *cfl_type |
85 | | #define CFL_hbd_TYPE uint16_t *cfl_type |
86 | | |
87 | | // Declare a size-specific wrapper for the size-generic function. The compiler |
88 | | // will inline the size generic function in here, the advantage is that the size |
89 | | // will be constant allowing for loop unrolling and other constant propagated |
90 | | // goodness. |
91 | | #define CFL_SUBSAMPLE(arch, sub, bd, width, height) \ |
92 | | void cfl_subsample_##bd##_##sub##_##width##x##height##_##arch( \ |
93 | 19.2M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ |
94 | 19.2M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ |
95 | 19.2M | output_q3, width, height); \ |
96 | 19.2M | } 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 | 93 | 4.46M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 4.46M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 4.46M | output_q3, width, height); \ | 96 | 4.46M | } |
cfl_subsample_lbd_420_8x8_ssse3 Line | Count | Source | 93 | 1.31M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.31M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.31M | output_q3, width, height); \ | 96 | 1.31M | } |
cfl_subsample_lbd_420_16x16_ssse3 Line | Count | Source | 93 | 422k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 422k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 422k | output_q3, width, height); \ | 96 | 422k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x32_ssse3 cfl_subsample_lbd_420_4x8_ssse3 Line | Count | Source | 93 | 4.72M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 4.72M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 4.72M | output_q3, width, height); \ | 96 | 4.72M | } |
cfl_subsample_lbd_420_8x4_ssse3 Line | Count | Source | 93 | 3.09M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 3.09M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 3.09M | output_q3, width, height); \ | 96 | 3.09M | } |
cfl_subsample_lbd_420_8x16_ssse3 Line | Count | Source | 93 | 71.2k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 71.2k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 71.2k | output_q3, width, height); \ | 96 | 71.2k | } |
cfl_subsample_lbd_420_16x8_ssse3 Line | Count | Source | 93 | 155k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 155k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 155k | output_q3, width, height); \ | 96 | 155k | } |
cfl_subsample_lbd_420_16x32_ssse3 Line | Count | Source | 93 | 29.8k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 29.8k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 29.8k | output_q3, width, height); \ | 96 | 29.8k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x16_ssse3 cfl_subsample_lbd_420_4x16_ssse3 Line | Count | Source | 93 | 1.66M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.66M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.66M | output_q3, width, height); \ | 96 | 1.66M | } |
cfl_subsample_lbd_420_16x4_ssse3 Line | Count | Source | 93 | 1.38M | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.38M | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.38M | output_q3, width, height); \ | 96 | 1.38M | } |
cfl_subsample_lbd_420_8x32_ssse3 Line | Count | Source | 93 | 2.81k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.81k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.81k | output_q3, width, height); \ | 96 | 2.81k | } |
Unexecuted instantiation: cfl_subsample_lbd_420_32x8_ssse3 cfl_subsample_lbd_422_4x4_ssse3 Line | Count | Source | 93 | 1.65k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.65k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.65k | output_q3, width, height); \ | 96 | 1.65k | } |
cfl_subsample_lbd_422_8x8_ssse3 Line | Count | Source | 93 | 2.01k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.01k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.01k | output_q3, width, height); \ | 96 | 2.01k | } |
cfl_subsample_lbd_422_16x16_ssse3 Line | Count | Source | 93 | 337 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 337 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 337 | output_q3, width, height); \ | 96 | 337 | } |
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 | 93 | 507 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 507 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 507 | output_q3, width, height); \ | 96 | 507 | } |
Unexecuted instantiation: cfl_subsample_lbd_422_8x16_ssse3 cfl_subsample_lbd_422_16x8_ssse3 Line | Count | Source | 93 | 504 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 504 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 504 | output_q3, width, height); \ | 96 | 504 | } |
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 | 93 | 288 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 288 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 288 | output_q3, width, height); \ | 96 | 288 | } |
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 | 93 | 484 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 484 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 484 | output_q3, width, height); \ | 96 | 484 | } |
cfl_subsample_lbd_444_8x8_ssse3 Line | Count | Source | 93 | 1.31k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.31k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.31k | output_q3, width, height); \ | 96 | 1.31k | } |
cfl_subsample_lbd_444_16x16_ssse3 Line | Count | Source | 93 | 134 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 134 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 134 | output_q3, width, height); \ | 96 | 134 | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x32_ssse3 cfl_subsample_lbd_444_4x8_ssse3 Line | Count | Source | 93 | 117 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 117 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 117 | output_q3, width, height); \ | 96 | 117 | } |
cfl_subsample_lbd_444_8x4_ssse3 Line | Count | Source | 93 | 196 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 196 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 196 | output_q3, width, height); \ | 96 | 196 | } |
cfl_subsample_lbd_444_8x16_ssse3 Line | Count | Source | 93 | 93 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 93 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 93 | output_q3, width, height); \ | 96 | 93 | } |
cfl_subsample_lbd_444_16x8_ssse3 Line | Count | Source | 93 | 105 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 105 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 105 | output_q3, width, height); \ | 96 | 105 | } |
cfl_subsample_lbd_444_16x32_ssse3 Line | Count | Source | 93 | 55 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 55 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 55 | output_q3, width, height); \ | 96 | 55 | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x16_ssse3 cfl_subsample_lbd_444_4x16_ssse3 Line | Count | Source | 93 | 89 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 89 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 89 | output_q3, width, height); \ | 96 | 89 | } |
cfl_subsample_lbd_444_16x4_ssse3 Line | Count | Source | 93 | 68 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 68 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 68 | output_q3, width, height); \ | 96 | 68 | } |
cfl_subsample_lbd_444_8x32_ssse3 Line | Count | Source | 93 | 394 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 394 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 394 | output_q3, width, height); \ | 96 | 394 | } |
Unexecuted instantiation: cfl_subsample_lbd_444_32x8_ssse3 cfl_subsample_hbd_420_4x4_ssse3 Line | Count | Source | 93 | 369k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 369k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 369k | output_q3, width, height); \ | 96 | 369k | } |
cfl_subsample_hbd_420_8x8_ssse3 Line | Count | Source | 93 | 165k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 165k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 165k | output_q3, width, height); \ | 96 | 165k | } |
cfl_subsample_hbd_420_16x16_ssse3 Line | Count | Source | 93 | 81.4k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 81.4k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 81.4k | output_q3, width, height); \ | 96 | 81.4k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x32_ssse3 cfl_subsample_hbd_420_4x8_ssse3 Line | Count | Source | 93 | 357k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 357k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 357k | output_q3, width, height); \ | 96 | 357k | } |
cfl_subsample_hbd_420_8x4_ssse3 Line | Count | Source | 93 | 232k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 232k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 232k | output_q3, width, height); \ | 96 | 232k | } |
cfl_subsample_hbd_420_8x16_ssse3 Line | Count | Source | 93 | 7.25k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 7.25k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 7.25k | output_q3, width, height); \ | 96 | 7.25k | } |
cfl_subsample_hbd_420_16x8_ssse3 Line | Count | Source | 93 | 37.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 37.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 37.9k | output_q3, width, height); \ | 96 | 37.9k | } |
cfl_subsample_hbd_420_16x32_ssse3 Line | Count | Source | 93 | 2.64k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.64k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.64k | output_q3, width, height); \ | 96 | 2.64k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x16_ssse3 cfl_subsample_hbd_420_4x16_ssse3 Line | Count | Source | 93 | 141k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 141k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 141k | output_q3, width, height); \ | 96 | 141k | } |
cfl_subsample_hbd_420_16x4_ssse3 Line | Count | Source | 93 | 157k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 157k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 157k | output_q3, width, height); \ | 96 | 157k | } |
cfl_subsample_hbd_420_8x32_ssse3 Line | Count | Source | 93 | 1.30k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.30k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.30k | output_q3, width, height); \ | 96 | 1.30k | } |
Unexecuted instantiation: cfl_subsample_hbd_420_32x8_ssse3 cfl_subsample_hbd_422_4x4_ssse3 Line | Count | Source | 93 | 1.43k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.43k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.43k | output_q3, width, height); \ | 96 | 1.43k | } |
cfl_subsample_hbd_422_8x8_ssse3 Line | Count | Source | 93 | 2.77k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.77k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.77k | output_q3, width, height); \ | 96 | 2.77k | } |
cfl_subsample_hbd_422_16x16_ssse3 Line | Count | Source | 93 | 348 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 348 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 348 | output_q3, width, height); \ | 96 | 348 | } |
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 | 93 | 485 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 485 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 485 | output_q3, width, height); \ | 96 | 485 | } |
Unexecuted instantiation: cfl_subsample_hbd_422_8x16_ssse3 cfl_subsample_hbd_422_16x8_ssse3 Line | Count | Source | 93 | 420 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 420 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 420 | output_q3, width, height); \ | 96 | 420 | } |
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 | 93 | 494 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 494 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 494 | output_q3, width, height); \ | 96 | 494 | } |
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 | 93 | 39.4k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 39.4k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 39.4k | output_q3, width, height); \ | 96 | 39.4k | } |
cfl_subsample_hbd_444_8x8_ssse3 Line | Count | Source | 93 | 59.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 59.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 59.1k | output_q3, width, height); \ | 96 | 59.1k | } |
cfl_subsample_hbd_444_16x16_ssse3 Line | Count | Source | 93 | 6.39k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 6.39k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 6.39k | output_q3, width, height); \ | 96 | 6.39k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x32_ssse3 cfl_subsample_hbd_444_4x8_ssse3 Line | Count | Source | 93 | 6.28k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 6.28k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 6.28k | output_q3, width, height); \ | 96 | 6.28k | } |
cfl_subsample_hbd_444_8x4_ssse3 Line | Count | Source | 93 | 9.90k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 9.90k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 9.90k | output_q3, width, height); \ | 96 | 9.90k | } |
cfl_subsample_hbd_444_8x16_ssse3 Line | Count | Source | 93 | 4.48k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 4.48k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 4.48k | output_q3, width, height); \ | 96 | 4.48k | } |
cfl_subsample_hbd_444_16x8_ssse3 Line | Count | Source | 93 | 6.48k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 6.48k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 6.48k | output_q3, width, height); \ | 96 | 6.48k | } |
cfl_subsample_hbd_444_16x32_ssse3 Line | Count | Source | 93 | 1.57k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 1.57k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 1.57k | output_q3, width, height); \ | 96 | 1.57k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x16_ssse3 cfl_subsample_hbd_444_4x16_ssse3 Line | Count | Source | 93 | 2.46k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.46k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.46k | output_q3, width, height); \ | 96 | 2.46k | } |
cfl_subsample_hbd_444_16x4_ssse3 Line | Count | Source | 93 | 4.77k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 4.77k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 4.77k | output_q3, width, height); \ | 96 | 4.77k | } |
cfl_subsample_hbd_444_8x32_ssse3 Line | Count | Source | 93 | 2.32k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.32k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.32k | output_q3, width, height); \ | 96 | 2.32k | } |
Unexecuted instantiation: cfl_subsample_hbd_444_32x8_ssse3 cfl_subsample_lbd_420_32x32_avx2 Line | Count | Source | 93 | 66.4k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 66.4k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 66.4k | output_q3, width, height); \ | 96 | 66.4k | } |
cfl_subsample_lbd_420_32x16_avx2 Line | Count | Source | 93 | 42.1k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 42.1k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 42.1k | output_q3, width, height); \ | 96 | 42.1k | } |
cfl_subsample_lbd_420_32x8_avx2 Line | Count | Source | 93 | 17.5k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 17.5k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 17.5k | output_q3, width, height); \ | 96 | 17.5k | } |
cfl_subsample_lbd_422_32x32_avx2 Line | Count | Source | 93 | 242 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 242 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 242 | output_q3, width, height); \ | 96 | 242 | } |
cfl_subsample_lbd_422_32x16_avx2 Line | Count | Source | 93 | 302 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 302 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 302 | output_q3, width, height); \ | 96 | 302 | } |
cfl_subsample_lbd_422_32x8_avx2 Line | Count | Source | 93 | 447 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 447 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 447 | output_q3, width, height); \ | 96 | 447 | } |
cfl_subsample_lbd_444_32x32_avx2 Line | Count | Source | 93 | 102 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 102 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 102 | output_q3, width, height); \ | 96 | 102 | } |
cfl_subsample_lbd_444_32x16_avx2 Line | Count | Source | 93 | 58 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 58 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 58 | output_q3, width, height); \ | 96 | 58 | } |
cfl_subsample_lbd_444_32x8_avx2 Line | Count | Source | 93 | 63 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 63 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 63 | output_q3, width, height); \ | 96 | 63 | } |
cfl_subsample_hbd_420_32x32_avx2 Line | Count | Source | 93 | 7.78k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 7.78k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 7.78k | output_q3, width, height); \ | 96 | 7.78k | } |
cfl_subsample_hbd_420_32x16_avx2 Line | Count | Source | 93 | 12.9k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 12.9k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 12.9k | output_q3, width, height); \ | 96 | 12.9k | } |
cfl_subsample_hbd_420_32x8_avx2 Line | Count | Source | 93 | 8.94k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 8.94k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 8.94k | output_q3, width, height); \ | 96 | 8.94k | } |
cfl_subsample_hbd_422_32x32_avx2 Line | Count | Source | 93 | 477 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 477 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 477 | output_q3, width, height); \ | 96 | 477 | } |
cfl_subsample_hbd_422_32x16_avx2 Line | Count | Source | 93 | 235 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 235 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 235 | output_q3, width, height); \ | 96 | 235 | } |
cfl_subsample_hbd_422_32x8_avx2 Line | Count | Source | 93 | 377 | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 377 | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 377 | output_q3, width, height); \ | 96 | 377 | } |
cfl_subsample_hbd_444_32x32_avx2 Line | Count | Source | 93 | 4.04k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 4.04k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 4.04k | output_q3, width, height); \ | 96 | 4.04k | } |
cfl_subsample_hbd_444_32x16_avx2 Line | Count | Source | 93 | 3.74k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 3.74k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 3.74k | output_q3, width, height); \ | 96 | 3.74k | } |
cfl_subsample_hbd_444_32x8_avx2 Line | Count | Source | 93 | 2.67k | const CFL_##bd##_TYPE, int input_stride, uint16_t *output_q3) { \ | 94 | 2.67k | cfl_luma_subsampling_##sub##_##bd##_##arch(cfl_type, input_stride, \ | 95 | 2.67k | output_q3, width, height); \ | 96 | 2.67k | } |
|
97 | | |
98 | | // Declare size-specific wrappers for all valid CfL sizes. |
99 | | #define CFL_SUBSAMPLE_FUNCTIONS(arch, sub, bd) \ |
100 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 4) \ |
101 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 8) \ |
102 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 16) \ |
103 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 32) \ |
104 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 8) \ |
105 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 4) \ |
106 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 16) \ |
107 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 8) \ |
108 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 32) \ |
109 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 16) \ |
110 | | CFL_SUBSAMPLE(arch, sub, bd, 4, 16) \ |
111 | | CFL_SUBSAMPLE(arch, sub, bd, 16, 4) \ |
112 | | CFL_SUBSAMPLE(arch, sub, bd, 8, 32) \ |
113 | | CFL_SUBSAMPLE(arch, sub, bd, 32, 8) \ |
114 | | cfl_subsample_##bd##_fn cfl_get_luma_subsampling_##sub##_##bd##_##arch( \ |
115 | 0 | TX_SIZE tx_size) { \ |
116 | 0 | CFL_SUBSAMPLE_FUNCTION_ARRAY(arch, sub, bd) \ |
117 | 0 | return subfn_##sub[tx_size]; \ |
118 | 0 | } Unexecuted instantiation: cfl_get_luma_subsampling_420_lbd_c Unexecuted instantiation: cfl_get_luma_subsampling_422_lbd_c Unexecuted instantiation: cfl_get_luma_subsampling_444_lbd_c Unexecuted instantiation: cfl_get_luma_subsampling_420_hbd_c Unexecuted instantiation: cfl_get_luma_subsampling_422_hbd_c Unexecuted instantiation: cfl_get_luma_subsampling_444_hbd_c Unexecuted instantiation: cfl_get_luma_subsampling_420_lbd_ssse3 Unexecuted instantiation: cfl_get_luma_subsampling_422_lbd_ssse3 Unexecuted instantiation: cfl_get_luma_subsampling_444_lbd_ssse3 Unexecuted instantiation: cfl_get_luma_subsampling_420_hbd_ssse3 Unexecuted instantiation: cfl_get_luma_subsampling_422_hbd_ssse3 Unexecuted instantiation: cfl_get_luma_subsampling_444_hbd_ssse3 |
119 | | |
120 | | // Declare an architecture-specific array of function pointers for size-specific |
121 | | // wrappers. |
122 | | #define CFL_SUBSAMPLE_FUNCTION_ARRAY(arch, sub, bd) \ |
123 | 0 | static const cfl_subsample_##bd##_fn subfn_##sub[TX_SIZES_ALL] = { \ |
124 | 0 | cfl_subsample_##bd##_##sub##_4x4_##arch, /* 4x4 */ \ |
125 | 0 | cfl_subsample_##bd##_##sub##_8x8_##arch, /* 8x8 */ \ |
126 | 0 | cfl_subsample_##bd##_##sub##_16x16_##arch, /* 16x16 */ \ |
127 | 0 | cfl_subsample_##bd##_##sub##_32x32_##arch, /* 32x32 */ \ |
128 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
129 | 0 | cfl_subsample_##bd##_##sub##_4x8_##arch, /* 4x8 */ \ |
130 | 0 | cfl_subsample_##bd##_##sub##_8x4_##arch, /* 8x4 */ \ |
131 | 0 | cfl_subsample_##bd##_##sub##_8x16_##arch, /* 8x16 */ \ |
132 | 0 | cfl_subsample_##bd##_##sub##_16x8_##arch, /* 16x8 */ \ |
133 | 0 | cfl_subsample_##bd##_##sub##_16x32_##arch, /* 16x32 */ \ |
134 | 0 | cfl_subsample_##bd##_##sub##_32x16_##arch, /* 32x16 */ \ |
135 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
136 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
137 | 0 | cfl_subsample_##bd##_##sub##_4x16_##arch, /* 4x16 */ \ |
138 | 0 | cfl_subsample_##bd##_##sub##_16x4_##arch, /* 16x4 */ \ |
139 | 0 | cfl_subsample_##bd##_##sub##_8x32_##arch, /* 8x32 */ \ |
140 | 0 | cfl_subsample_##bd##_##sub##_32x8_##arch, /* 32x8 */ \ |
141 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
142 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
143 | 0 | }; |
144 | | |
145 | | // The RTCD script does not support passing in an array, so we wrap it in this |
146 | | // function. |
147 | | #if CONFIG_AV1_HIGHBITDEPTH |
148 | | #define CFL_GET_SUBSAMPLE_FUNCTION(arch) \ |
149 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 420, lbd) \ |
150 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 422, lbd) \ |
151 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 444, lbd) \ |
152 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 420, hbd) \ |
153 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 422, hbd) \ |
154 | | CFL_SUBSAMPLE_FUNCTIONS(arch, 444, hbd) |
155 | | #else |
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 | | #endif |
161 | | |
162 | | // Declare a size-specific wrapper for the size-generic function. The compiler |
163 | | // will inline the size generic function in here, the advantage is that the size |
164 | | // will be constant allowing for loop unrolling and other constant propagated |
165 | | // goodness. |
166 | | #define CFL_SUB_AVG_X(arch, width, height, round_offset, num_pel_log2) \ |
167 | | void cfl_subtract_average_##width##x##height##_##arch(const uint16_t *src, \ |
168 | 2.21M | int16_t *dst) { \ |
169 | 2.21M | subtract_average_##arch(src, dst, width, height, round_offset, \ |
170 | 2.21M | num_pel_log2); \ |
171 | 2.21M | } 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 | 168 | 471k | int16_t *dst) { \ | 169 | 471k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 471k | num_pel_log2); \ | 171 | 471k | } |
cfl_subtract_average_4x8_sse2 Line | Count | Source | 168 | 520k | int16_t *dst) { \ | 169 | 520k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 520k | num_pel_log2); \ | 171 | 520k | } |
cfl_subtract_average_4x16_sse2 Line | Count | Source | 168 | 10.3k | int16_t *dst) { \ | 169 | 10.3k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 10.3k | num_pel_log2); \ | 171 | 10.3k | } |
cfl_subtract_average_8x4_sse2 Line | Count | Source | 168 | 308k | int16_t *dst) { \ | 169 | 308k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 308k | num_pel_log2); \ | 171 | 308k | } |
cfl_subtract_average_8x8_sse2 Line | Count | Source | 168 | 568k | int16_t *dst) { \ | 169 | 568k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 568k | num_pel_log2); \ | 171 | 568k | } |
cfl_subtract_average_8x16_sse2 Line | Count | Source | 168 | 43.4k | int16_t *dst) { \ | 169 | 43.4k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 43.4k | num_pel_log2); \ | 171 | 43.4k | } |
cfl_subtract_average_8x32_sse2 Line | Count | Source | 168 | 3.60k | int16_t *dst) { \ | 169 | 3.60k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 3.60k | num_pel_log2); \ | 171 | 3.60k | } |
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 | 168 | 65.3k | int16_t *dst) { \ | 169 | 65.3k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 65.3k | num_pel_log2); \ | 171 | 65.3k | } |
cfl_subtract_average_16x8_avx2 Line | Count | Source | 168 | 78.2k | int16_t *dst) { \ | 169 | 78.2k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 78.2k | num_pel_log2); \ | 171 | 78.2k | } |
cfl_subtract_average_16x16_avx2 Line | Count | Source | 168 | 124k | int16_t *dst) { \ | 169 | 124k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 124k | num_pel_log2); \ | 171 | 124k | } |
cfl_subtract_average_16x32_avx2 Line | Count | Source | 168 | 3.24k | int16_t *dst) { \ | 169 | 3.24k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 3.24k | num_pel_log2); \ | 171 | 3.24k | } |
cfl_subtract_average_32x8_avx2 Line | Count | Source | 168 | 3.58k | int16_t *dst) { \ | 169 | 3.58k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 3.58k | num_pel_log2); \ | 171 | 3.58k | } |
cfl_subtract_average_32x16_avx2 Line | Count | Source | 168 | 4.46k | int16_t *dst) { \ | 169 | 4.46k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 4.46k | num_pel_log2); \ | 171 | 4.46k | } |
cfl_subtract_average_32x32_avx2 Line | Count | Source | 168 | 5.34k | int16_t *dst) { \ | 169 | 5.34k | subtract_average_##arch(src, dst, width, height, round_offset, \ | 170 | 5.34k | num_pel_log2); \ | 171 | 5.34k | } |
|
172 | | |
173 | | // Declare size-specific wrappers for all valid CfL sizes. |
174 | | #define CFL_SUB_AVG_FN(arch) \ |
175 | | CFL_SUB_AVG_X(arch, 4, 4, 8, 4) \ |
176 | | CFL_SUB_AVG_X(arch, 4, 8, 16, 5) \ |
177 | | CFL_SUB_AVG_X(arch, 4, 16, 32, 6) \ |
178 | | CFL_SUB_AVG_X(arch, 8, 4, 16, 5) \ |
179 | | CFL_SUB_AVG_X(arch, 8, 8, 32, 6) \ |
180 | | CFL_SUB_AVG_X(arch, 8, 16, 64, 7) \ |
181 | | CFL_SUB_AVG_X(arch, 8, 32, 128, 8) \ |
182 | | CFL_SUB_AVG_X(arch, 16, 4, 32, 6) \ |
183 | | CFL_SUB_AVG_X(arch, 16, 8, 64, 7) \ |
184 | | CFL_SUB_AVG_X(arch, 16, 16, 128, 8) \ |
185 | | CFL_SUB_AVG_X(arch, 16, 32, 256, 9) \ |
186 | | CFL_SUB_AVG_X(arch, 32, 8, 128, 8) \ |
187 | | CFL_SUB_AVG_X(arch, 32, 16, 256, 9) \ |
188 | | CFL_SUB_AVG_X(arch, 32, 32, 512, 10) \ |
189 | | cfl_subtract_average_fn cfl_get_subtract_average_fn_##arch( \ |
190 | 0 | TX_SIZE tx_size) { \ |
191 | 0 | static const cfl_subtract_average_fn sub_avg[TX_SIZES_ALL] = { \ |
192 | 0 | cfl_subtract_average_4x4_##arch, /* 4x4 */ \ |
193 | 0 | cfl_subtract_average_8x8_##arch, /* 8x8 */ \ |
194 | 0 | cfl_subtract_average_16x16_##arch, /* 16x16 */ \ |
195 | 0 | cfl_subtract_average_32x32_##arch, /* 32x32 */ \ |
196 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
197 | 0 | cfl_subtract_average_4x8_##arch, /* 4x8 */ \ |
198 | 0 | cfl_subtract_average_8x4_##arch, /* 8x4 */ \ |
199 | 0 | cfl_subtract_average_8x16_##arch, /* 8x16 */ \ |
200 | 0 | cfl_subtract_average_16x8_##arch, /* 16x8 */ \ |
201 | 0 | cfl_subtract_average_16x32_##arch, /* 16x32 */ \ |
202 | 0 | cfl_subtract_average_32x16_##arch, /* 32x16 */ \ |
203 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
204 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
205 | 0 | cfl_subtract_average_4x16_##arch, /* 4x16 (invalid CFL size) */ \ |
206 | 0 | cfl_subtract_average_16x4_##arch, /* 16x4 (invalid CFL size) */ \ |
207 | 0 | cfl_subtract_average_8x32_##arch, /* 8x32 (invalid CFL size) */ \ |
208 | 0 | cfl_subtract_average_32x8_##arch, /* 32x8 (invalid CFL size) */ \ |
209 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
210 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
211 | 0 | }; \ |
212 | 0 | /* Modulo TX_SIZES_ALL to ensure that an attacker won't be able to */ \ |
213 | 0 | /* index the function pointer array out of bounds. */ \ |
214 | 0 | return sub_avg[tx_size % TX_SIZES_ALL]; \ |
215 | 0 | } Unexecuted instantiation: cfl_get_subtract_average_fn_c Unexecuted instantiation: cfl_get_subtract_average_fn_sse2 |
216 | | |
217 | | // For VSX SIMD optimization, the C versions of width == 4 subtract are |
218 | | // faster than the VSX. As such, the VSX code calls the C versions. |
219 | | void cfl_subtract_average_4x4_c(const uint16_t *src, int16_t *dst); |
220 | | void cfl_subtract_average_4x8_c(const uint16_t *src, int16_t *dst); |
221 | | void cfl_subtract_average_4x16_c(const uint16_t *src, int16_t *dst); |
222 | | |
223 | | #define CFL_PREDICT_lbd(arch, width, height) \ |
224 | | void cfl_predict_lbd_##width##x##height##_##arch( \ |
225 | | const int16_t *pred_buf_q3, uint8_t *dst, int dst_stride, \ |
226 | 3.75M | int alpha_q3) { \ |
227 | 3.75M | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ |
228 | 3.75M | height); \ |
229 | 3.75M | } 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 | 226 | 836k | int alpha_q3) { \ | 227 | 836k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 836k | height); \ | 229 | 836k | } |
cfl_predict_lbd_4x8_ssse3 Line | Count | Source | 226 | 948k | int alpha_q3) { \ | 227 | 948k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 948k | height); \ | 229 | 948k | } |
cfl_predict_lbd_4x16_ssse3 Line | Count | Source | 226 | 9.84k | int alpha_q3) { \ | 227 | 9.84k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 9.84k | height); \ | 229 | 9.84k | } |
cfl_predict_lbd_8x4_ssse3 Line | Count | Source | 226 | 523k | int alpha_q3) { \ | 227 | 523k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 523k | height); \ | 229 | 523k | } |
cfl_predict_lbd_8x8_ssse3 Line | Count | Source | 226 | 993k | int alpha_q3) { \ | 227 | 993k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 993k | height); \ | 229 | 993k | } |
cfl_predict_lbd_8x16_ssse3 Line | Count | Source | 226 | 64.6k | int alpha_q3) { \ | 227 | 64.6k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 64.6k | height); \ | 229 | 64.6k | } |
cfl_predict_lbd_8x32_ssse3 Line | Count | Source | 226 | 810 | int alpha_q3) { \ | 227 | 810 | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 810 | height); \ | 229 | 810 | } |
cfl_predict_lbd_16x4_ssse3 Line | Count | Source | 226 | 77.9k | int alpha_q3) { \ | 227 | 77.9k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 77.9k | height); \ | 229 | 77.9k | } |
cfl_predict_lbd_16x8_ssse3 Line | Count | Source | 226 | 99.9k | int alpha_q3) { \ | 227 | 99.9k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 99.9k | height); \ | 229 | 99.9k | } |
cfl_predict_lbd_16x16_ssse3 Line | Count | Source | 226 | 197k | int alpha_q3) { \ | 227 | 197k | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 197k | height); \ | 229 | 197k | } |
cfl_predict_lbd_16x32_ssse3 Line | Count | Source | 226 | 794 | int alpha_q3) { \ | 227 | 794 | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 794 | height); \ | 229 | 794 | } |
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 | 226 | 143 | int alpha_q3) { \ | 227 | 143 | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 143 | height); \ | 229 | 143 | } |
cfl_predict_lbd_32x16_avx2 Line | Count | Source | 226 | 126 | int alpha_q3) { \ | 227 | 126 | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 126 | height); \ | 229 | 126 | } |
cfl_predict_lbd_32x32_avx2 Line | Count | Source | 226 | 357 | int alpha_q3) { \ | 227 | 357 | cfl_predict_lbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, width, \ | 228 | 357 | height); \ | 229 | 357 | } |
|
230 | | |
231 | | #if CONFIG_AV1_HIGHBITDEPTH |
232 | | #define CFL_PREDICT_hbd(arch, width, height) \ |
233 | | void cfl_predict_hbd_##width##x##height##_##arch( \ |
234 | | const int16_t *pred_buf_q3, uint16_t *dst, int dst_stride, int alpha_q3, \ |
235 | 669k | int bd) { \ |
236 | 669k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ |
237 | 669k | height); \ |
238 | 669k | } 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 | 235 | 107k | int bd) { \ | 236 | 107k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 107k | height); \ | 238 | 107k | } |
cfl_predict_hbd_4x8_ssse3 Line | Count | Source | 235 | 93.0k | int bd) { \ | 236 | 93.0k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 93.0k | height); \ | 238 | 93.0k | } |
cfl_predict_hbd_4x16_ssse3 Line | Count | Source | 235 | 10.9k | int bd) { \ | 236 | 10.9k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 10.9k | height); \ | 238 | 10.9k | } |
cfl_predict_hbd_8x4_ssse3 Line | Count | Source | 235 | 92.8k | int bd) { \ | 236 | 92.8k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 92.8k | height); \ | 238 | 92.8k | } |
cfl_predict_hbd_8x8_ssse3 Line | Count | Source | 235 | 143k | int bd) { \ | 236 | 143k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 143k | height); \ | 238 | 143k | } |
cfl_predict_hbd_8x16_ssse3 Line | Count | Source | 235 | 22.2k | int bd) { \ | 236 | 22.2k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 22.2k | height); \ | 238 | 22.2k | } |
cfl_predict_hbd_8x32_ssse3 Line | Count | Source | 235 | 6.40k | int bd) { \ | 236 | 6.40k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 6.40k | height); \ | 238 | 6.40k | } |
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 | 235 | 52.6k | int bd) { \ | 236 | 52.6k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 52.6k | height); \ | 238 | 52.6k | } |
cfl_predict_hbd_16x8_avx2 Line | Count | Source | 235 | 56.6k | int bd) { \ | 236 | 56.6k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 56.6k | height); \ | 238 | 56.6k | } |
cfl_predict_hbd_16x16_avx2 Line | Count | Source | 235 | 52.2k | int bd) { \ | 236 | 52.2k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 52.2k | height); \ | 238 | 52.2k | } |
cfl_predict_hbd_16x32_avx2 Line | Count | Source | 235 | 5.68k | int bd) { \ | 236 | 5.68k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 5.68k | height); \ | 238 | 5.68k | } |
cfl_predict_hbd_32x8_avx2 Line | Count | Source | 235 | 7.02k | int bd) { \ | 236 | 7.02k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 7.02k | height); \ | 238 | 7.02k | } |
cfl_predict_hbd_32x16_avx2 Line | Count | Source | 235 | 8.80k | int bd) { \ | 236 | 8.80k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 8.80k | height); \ | 238 | 8.80k | } |
cfl_predict_hbd_32x32_avx2 Line | Count | Source | 235 | 10.3k | int bd) { \ | 236 | 10.3k | cfl_predict_hbd_##arch(pred_buf_q3, dst, dst_stride, alpha_q3, bd, width, \ | 237 | 10.3k | height); \ | 238 | 10.3k | } |
|
239 | | #endif |
240 | | |
241 | | // This wrapper exists because clang format does not like calling macros with |
242 | | // lowercase letters. |
243 | | #define CFL_PREDICT_X(arch, width, height, bd) \ |
244 | | CFL_PREDICT_##bd(arch, width, height) |
245 | | |
246 | | #define CFL_PREDICT_FN(arch, bd) \ |
247 | | CFL_PREDICT_X(arch, 4, 4, bd) \ |
248 | | CFL_PREDICT_X(arch, 4, 8, bd) \ |
249 | | CFL_PREDICT_X(arch, 4, 16, bd) \ |
250 | | CFL_PREDICT_X(arch, 8, 4, bd) \ |
251 | | CFL_PREDICT_X(arch, 8, 8, bd) \ |
252 | | CFL_PREDICT_X(arch, 8, 16, bd) \ |
253 | | CFL_PREDICT_X(arch, 8, 32, bd) \ |
254 | | CFL_PREDICT_X(arch, 16, 4, bd) \ |
255 | | CFL_PREDICT_X(arch, 16, 8, bd) \ |
256 | | CFL_PREDICT_X(arch, 16, 16, bd) \ |
257 | | CFL_PREDICT_X(arch, 16, 32, bd) \ |
258 | | CFL_PREDICT_X(arch, 32, 8, bd) \ |
259 | | CFL_PREDICT_X(arch, 32, 16, bd) \ |
260 | | CFL_PREDICT_X(arch, 32, 32, bd) \ |
261 | 0 | cfl_predict_##bd##_fn cfl_get_predict_##bd##_fn_##arch(TX_SIZE tx_size) { \ |
262 | 0 | static const cfl_predict_##bd##_fn pred[TX_SIZES_ALL] = { \ |
263 | 0 | cfl_predict_##bd##_4x4_##arch, /* 4x4 */ \ |
264 | 0 | cfl_predict_##bd##_8x8_##arch, /* 8x8 */ \ |
265 | 0 | cfl_predict_##bd##_16x16_##arch, /* 16x16 */ \ |
266 | 0 | cfl_predict_##bd##_32x32_##arch, /* 32x32 */ \ |
267 | 0 | NULL, /* 64x64 (invalid CFL size) */ \ |
268 | 0 | cfl_predict_##bd##_4x8_##arch, /* 4x8 */ \ |
269 | 0 | cfl_predict_##bd##_8x4_##arch, /* 8x4 */ \ |
270 | 0 | cfl_predict_##bd##_8x16_##arch, /* 8x16 */ \ |
271 | 0 | cfl_predict_##bd##_16x8_##arch, /* 16x8 */ \ |
272 | 0 | cfl_predict_##bd##_16x32_##arch, /* 16x32 */ \ |
273 | 0 | cfl_predict_##bd##_32x16_##arch, /* 32x16 */ \ |
274 | 0 | NULL, /* 32x64 (invalid CFL size) */ \ |
275 | 0 | NULL, /* 64x32 (invalid CFL size) */ \ |
276 | 0 | cfl_predict_##bd##_4x16_##arch, /* 4x16 */ \ |
277 | 0 | cfl_predict_##bd##_16x4_##arch, /* 16x4 */ \ |
278 | 0 | cfl_predict_##bd##_8x32_##arch, /* 8x32 */ \ |
279 | 0 | cfl_predict_##bd##_32x8_##arch, /* 32x8 */ \ |
280 | 0 | NULL, /* 16x64 (invalid CFL size) */ \ |
281 | 0 | NULL, /* 64x16 (invalid CFL size) */ \ |
282 | 0 | }; \ |
283 | 0 | /* Modulo TX_SIZES_ALL to ensure that an attacker won't be able to */ \ |
284 | 0 | /* index the function pointer array out of bounds. */ \ |
285 | 0 | return pred[tx_size % TX_SIZES_ALL]; \ |
286 | 0 | } Unexecuted instantiation: cfl_get_predict_lbd_fn_c Unexecuted instantiation: cfl_get_predict_hbd_fn_c Unexecuted instantiation: cfl_get_predict_lbd_fn_ssse3 Unexecuted instantiation: cfl_get_predict_hbd_fn_ssse3 |
287 | | |
288 | | #endif // AOM_AV1_COMMON_CFL_H_ |