1 /*
2 *
3 * This file is part of FFmpeg.
4 *
5 * FFmpeg is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU Lesser General Public
7 * License as published by the Free Software Foundation; either
8 * version 2.1 of the License, or (at your option) any later version.
9 *
10 * FFmpeg is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
14 *
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with FFmpeg; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18 */
19
20 #include "config.h"
21
22 #include "libavutil/attributes.h"
23 #include "libavutil/cpu.h"
24 #include "libavutil/x86/cpu.h"
25 #include "libavfilter/gblur.h"
26
27 void ff_horiz_slice_sse4(float *ptr, int width, int height, int steps, float nu, float bscale, float *localbuf);
28 void ff_horiz_slice_avx2(float *ptr, int width, int height, int steps, float nu, float bscale, float *localbuf);
29 void ff_horiz_slice_avx512(float *ptr, int width, int height, int steps, float nu, float bscale, float *localbuf);
30
31 void ff_postscale_slice_sse(float *ptr, int length, float postscale, float min, float max);
32 void ff_postscale_slice_avx2(float *ptr, int length, float postscale, float min, float max);
33 void ff_postscale_slice_avx512(float *ptr, int length, float postscale, float min, float max);
34
35 void ff_verti_slice_avx2(float *buffer, int width, int height, int column_begin, int column_end,
36 int steps, float nu, float bscale);
37 void ff_verti_slice_avx512(float *buffer, int width, int height, int column_begin, int column_end,
38 int steps, float nu, float bscale);
39
ff_gblur_init_x86(GBlurContext * s)40 av_cold void ff_gblur_init_x86(GBlurContext *s)
41 {
42 int cpu_flags = av_get_cpu_flags();
43
44 if (EXTERNAL_SSE(cpu_flags)) {
45 s->postscale_slice = ff_postscale_slice_sse;
46 }
47 if (EXTERNAL_AVX2_FAST(cpu_flags)) {
48 s->postscale_slice = ff_postscale_slice_avx2;
49 }
50 #if ARCH_X86_64
51 if (EXTERNAL_SSE4(cpu_flags)) {
52 s->horiz_slice = ff_horiz_slice_sse4;
53 }
54 if (EXTERNAL_AVX2(cpu_flags)) {
55 s->verti_slice = ff_verti_slice_avx2;
56 }
57 if (EXTERNAL_AVX512(cpu_flags)) {
58 s->postscale_slice = ff_postscale_slice_avx512;
59 s->verti_slice = ff_verti_slice_avx512;
60 }
61 if (EXTERNAL_AVX2(cpu_flags)) {
62 s->stride = EXTERNAL_AVX512(cpu_flags) ? 16 : 8;
63 s->localbuf = av_malloc(s->stride * sizeof(float) * s->planewidth[0] * s->planeheight[0]);
64 if (!s->localbuf)
65 return;
66
67 s->horiz_slice = ff_horiz_slice_avx2;
68 if (EXTERNAL_AVX512(cpu_flags)) {
69 s->horiz_slice = ff_horiz_slice_avx512;
70 }
71 }
72 #endif
73 }
74