Line | Branch | Exec | Source |
---|---|---|---|
1 | /* | ||
2 | * Copyright (c) 2025 Niklas Haas | ||
3 | * | ||
4 | * This file is part of FFmpeg. | ||
5 | * | ||
6 | * FFmpeg is free software; you can redistribute it and/or | ||
7 | * modify it under the terms of the GNU Lesser General Public | ||
8 | * License as published by the Free Software Foundation; either | ||
9 | * version 2.1 of the License, or (at your option) any later version. | ||
10 | * | ||
11 | * FFmpeg is distributed in the hope that it will be useful, | ||
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
14 | * Lesser General Public License for more details. | ||
15 | * | ||
16 | * You should have received a copy of the GNU Lesser General Public | ||
17 | * License along with FFmpeg; if not, write to the Free Software | ||
18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
19 | */ | ||
20 | |||
21 | #include "libavutil/attributes.h" | ||
22 | #include "libavutil/x86/cpu.h" | ||
23 | #include "libavfilter/vf_colordetect.h" | ||
24 | |||
25 | #define DETECT_RANGE_FUNC(FUNC_NAME, ASM_FUNC_NAME, C_FUNC_NAME, SHIFT, MMSIZE) \ | ||
26 | int ASM_FUNC_NAME(const uint8_t *src, ptrdiff_t stride, \ | ||
27 | ptrdiff_t width, ptrdiff_t height, int min, int max); \ | ||
28 | \ | ||
29 | static int FUNC_NAME(const uint8_t *src, ptrdiff_t stride, \ | ||
30 | ptrdiff_t width, ptrdiff_t height, int min, int max) \ | ||
31 | { \ | ||
32 | ptrdiff_t bytes = (width << SHIFT) & ~(MMSIZE - 1); \ | ||
33 | int ret = ASM_FUNC_NAME(src, stride, bytes, height, min, max); \ | ||
34 | if (ret) \ | ||
35 | return ret; \ | ||
36 | \ | ||
37 | return C_FUNC_NAME(src + bytes, stride, width - (bytes >> SHIFT), \ | ||
38 | height, min, max); \ | ||
39 | } | ||
40 | |||
41 | #define DETECT_ALPHA_FUNC(FUNC_NAME, ASM_FUNC_NAME, C_FUNC_NAME, SHIFT, MMSIZE) \ | ||
42 | int ASM_FUNC_NAME(const uint8_t *color, ptrdiff_t color_stride, \ | ||
43 | const uint8_t *alpha, ptrdiff_t alpha_stride, \ | ||
44 | ptrdiff_t width, ptrdiff_t height, int p, int q, int k); \ | ||
45 | \ | ||
46 | static int FUNC_NAME(const uint8_t *color, ptrdiff_t color_stride, \ | ||
47 | const uint8_t *alpha, ptrdiff_t alpha_stride, \ | ||
48 | ptrdiff_t width, ptrdiff_t height, int p, int q, int k) \ | ||
49 | { \ | ||
50 | ptrdiff_t bytes = (width << SHIFT) & ~(MMSIZE - 1); \ | ||
51 | int ret = ASM_FUNC_NAME(color, color_stride, alpha, alpha_stride, \ | ||
52 | bytes, height, p, q, k); \ | ||
53 | if (ret) \ | ||
54 | return ret; \ | ||
55 | \ | ||
56 | return C_FUNC_NAME(color + bytes, color_stride, alpha + bytes, alpha_stride,\ | ||
57 | width - (bytes >> SHIFT), height, p, q, k); \ | ||
58 | } | ||
59 | |||
60 | #if HAVE_X86ASM | ||
61 | #if HAVE_AVX512ICL_EXTERNAL | ||
62 | ✗ | DETECT_RANGE_FUNC(detect_range_avx512icl, ff_detect_rangeb_avx512icl, ff_detect_range_c, 0, 64) | |
63 | ✗ | DETECT_RANGE_FUNC(detect_range16_avx512icl, ff_detect_rangew_avx512icl, ff_detect_range16_c, 1, 64) | |
64 | ✗ | DETECT_ALPHA_FUNC(detect_alpha_full_avx512icl, ff_detect_alphab_full_avx512icl, ff_detect_alpha_full_c, 0, 64) | |
65 | ✗ | DETECT_ALPHA_FUNC(detect_alpha16_full_avx512icl, ff_detect_alphaw_full_avx512icl, ff_detect_alpha16_full_c, 1, 64) | |
66 | ✗ | DETECT_ALPHA_FUNC(detect_alpha_limited_avx512icl, ff_detect_alphab_limited_avx512icl, ff_detect_alpha_limited_c, 0, 64) | |
67 | ✗ | DETECT_ALPHA_FUNC(detect_alpha16_limited_avx512icl, ff_detect_alphaw_limited_avx512icl, ff_detect_alpha16_limited_c, 1, 64) | |
68 | #endif | ||
69 | #if HAVE_AVX2_EXTERNAL | ||
70 |
2/2✓ Branch 1 taken 3 times.
✓ Branch 2 taken 13 times.
|
16 | DETECT_RANGE_FUNC(detect_range_avx2, ff_detect_rangeb_avx2, ff_detect_range_c, 0, 32) |
71 |
2/2✓ Branch 1 taken 1 times.
✓ Branch 2 taken 15 times.
|
16 | DETECT_RANGE_FUNC(detect_range16_avx2, ff_detect_rangew_avx2, ff_detect_range16_c, 1, 32) |
72 |
2/2✓ Branch 1 taken 7 times.
✓ Branch 2 taken 9 times.
|
16 | DETECT_ALPHA_FUNC(detect_alpha_full_avx2, ff_detect_alphab_full_avx2, ff_detect_alpha_full_c, 0, 32) |
73 |
2/2✓ Branch 1 taken 6 times.
✓ Branch 2 taken 10 times.
|
16 | DETECT_ALPHA_FUNC(detect_alpha16_full_avx2, ff_detect_alphaw_full_avx2, ff_detect_alpha16_full_c, 1, 32) |
74 |
2/2✓ Branch 1 taken 3 times.
✓ Branch 2 taken 13 times.
|
16 | DETECT_ALPHA_FUNC(detect_alpha_limited_avx2, ff_detect_alphab_limited_avx2, ff_detect_alpha_limited_c, 0, 32) |
75 |
2/2✓ Branch 1 taken 1 times.
✓ Branch 2 taken 15 times.
|
16 | DETECT_ALPHA_FUNC(detect_alpha16_limited_avx2, ff_detect_alphaw_limited_avx2, ff_detect_alpha16_limited_c, 1, 32) |
76 | #endif | ||
77 | #endif | ||
78 | |||
79 | 78 | av_cold void ff_color_detect_dsp_init_x86(FFColorDetectDSPContext *dsp, int depth, | |
80 | enum AVColorRange color_range) | ||
81 | { | ||
82 | #if HAVE_X86ASM | ||
83 | 78 | int cpu_flags = av_get_cpu_flags(); | |
84 | #if HAVE_AVX2_EXTERNAL | ||
85 |
3/4✓ Branch 0 taken 6 times.
✓ Branch 1 taken 72 times.
✓ Branch 2 taken 6 times.
✗ Branch 3 not taken.
|
78 | if (EXTERNAL_AVX2_FAST(cpu_flags)) { |
86 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | dsp->detect_range = depth > 8 ? detect_range16_avx2 : detect_range_avx2; |
87 |
2/2✓ Branch 0 taken 2 times.
✓ Branch 1 taken 4 times.
|
6 | if (color_range == AVCOL_RANGE_JPEG) { |
88 |
2/2✓ Branch 0 taken 1 times.
✓ Branch 1 taken 1 times.
|
2 | dsp->detect_alpha = depth > 8 ? detect_alpha16_full_avx2 : detect_alpha_full_avx2; |
89 | } else { | ||
90 |
2/2✓ Branch 0 taken 2 times.
✓ Branch 1 taken 2 times.
|
4 | dsp->detect_alpha = depth > 8 ? detect_alpha16_limited_avx2 : detect_alpha_limited_avx2; |
91 | } | ||
92 | } | ||
93 | #endif | ||
94 | #if HAVE_AVX512ICL_EXTERNAL | ||
95 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 78 times.
|
78 | if (EXTERNAL_AVX512ICL(cpu_flags)) { |
96 | ✗ | dsp->detect_range = depth > 8 ? detect_range16_avx512icl : detect_range_avx512icl; | |
97 | ✗ | if (color_range == AVCOL_RANGE_JPEG) { | |
98 | ✗ | dsp->detect_alpha = depth > 8 ? detect_alpha16_full_avx512icl : detect_alpha_full_avx512icl; | |
99 | } else { | ||
100 | ✗ | dsp->detect_alpha = depth > 8 ? detect_alpha16_limited_avx512icl : detect_alpha_limited_avx512icl; | |
101 | } | ||
102 | } | ||
103 | #endif | ||
104 | #endif | ||
105 | 78 | } | |
106 |