2 * Format Conversion Utils
3 * Copyright (c) 2000, 2001 Fabrice Bellard
4 * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 * MMX optimization by Nick Kurshev <nickols_k@mail.ru>
25 #include "libavutil/attributes.h"
26 #include "libavutil/cpu.h"
27 #include "libavutil/x86/asm.h"
28 #include "libavutil/x86/cpu.h"
29 #include "libavcodec/fmtconvert.h"
30 #include "libavcodec/dsputil.h"
34 void ff_int32_to_float_fmul_scalar_sse (float *dst, const int *src, float mul, int len);
35 void ff_int32_to_float_fmul_scalar_sse2(float *dst, const int *src, float mul, int len);
37 void ff_float_to_int16_3dnow(int16_t *dst, const float *src, long len);
38 void ff_float_to_int16_sse (int16_t *dst, const float *src, long len);
39 void ff_float_to_int16_sse2 (int16_t *dst, const float *src, long len);
41 void ff_float_to_int16_step_3dnow(int16_t *dst, const float *src, long len, long step);
42 void ff_float_to_int16_step_sse (int16_t *dst, const float *src, long len, long step);
43 void ff_float_to_int16_step_sse2 (int16_t *dst, const float *src, long len, long step);
45 void ff_float_to_int16_interleave2_3dnow(int16_t *dst, const float **src, long len);
46 void ff_float_to_int16_interleave2_sse (int16_t *dst, const float **src, long len);
47 void ff_float_to_int16_interleave2_sse2 (int16_t *dst, const float **src, long len);
49 void ff_float_to_int16_interleave6_sse(int16_t *dst, const float **src, int len);
50 void ff_float_to_int16_interleave6_3dnow(int16_t *dst, const float **src, int len);
51 void ff_float_to_int16_interleave6_3dnowext(int16_t *dst, const float **src, int len);
53 #define ff_float_to_int16_interleave6_sse2 ff_float_to_int16_interleave6_sse
55 #define FLOAT_TO_INT16_INTERLEAVE(cpu) \
56 /* gcc pessimizes register allocation if this is in the same function as float_to_int16_interleave_sse2*/\
57 static av_noinline void float_to_int16_interleave_misc_##cpu(int16_t *dst, const float **src, long len, int channels){\
59 for(c=0; c<channels; c++){\
60 ff_float_to_int16_step_##cpu(dst+c, src[c], len, channels);\
64 static void float_to_int16_interleave_##cpu(int16_t *dst, const float **src, long len, int channels){\
66 ff_float_to_int16_##cpu(dst, src[0], len);\
67 else if(channels==2){\
68 ff_float_to_int16_interleave2_##cpu(dst, src, len);\
69 }else if(channels==6){\
70 ff_float_to_int16_interleave6_##cpu(dst, src, len);\
72 float_to_int16_interleave_misc_##cpu(dst, src, len, channels);\
75 FLOAT_TO_INT16_INTERLEAVE(3dnow)
76 FLOAT_TO_INT16_INTERLEAVE(sse)
77 FLOAT_TO_INT16_INTERLEAVE(sse2)
79 static void float_to_int16_interleave_3dnowext(int16_t *dst, const float **src,
80 long len, int channels)
83 ff_float_to_int16_interleave6_3dnowext(dst, src, len);
85 float_to_int16_interleave_3dnow(dst, src, len, channels);
88 void ff_float_interleave2_mmx(float *dst, const float **src, unsigned int len);
89 void ff_float_interleave2_sse(float *dst, const float **src, unsigned int len);
91 void ff_float_interleave6_mmx(float *dst, const float **src, unsigned int len);
92 void ff_float_interleave6_sse(float *dst, const float **src, unsigned int len);
94 static void float_interleave_mmx(float *dst, const float **src,
95 unsigned int len, int channels)
98 ff_float_interleave2_mmx(dst, src, len);
99 } else if (channels == 6)
100 ff_float_interleave6_mmx(dst, src, len);
102 ff_float_interleave_c(dst, src, len, channels);
105 static void float_interleave_sse(float *dst, const float **src,
106 unsigned int len, int channels)
109 ff_float_interleave2_sse(dst, src, len);
110 } else if (channels == 6)
111 ff_float_interleave6_sse(dst, src, len);
113 ff_float_interleave_c(dst, src, len, channels);
115 #endif /* HAVE_YASM */
117 av_cold void ff_fmt_convert_init_x86(FmtConvertContext *c, AVCodecContext *avctx)
120 int mm_flags = av_get_cpu_flags();
122 if (EXTERNAL_MMX(mm_flags)) {
123 c->float_interleave = float_interleave_mmx;
125 if (EXTERNAL_AMD3DNOW(mm_flags)) {
126 if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
127 c->float_to_int16 = ff_float_to_int16_3dnow;
128 c->float_to_int16_interleave = float_to_int16_interleave_3dnow;
131 if (EXTERNAL_AMD3DNOWEXT(mm_flags)) {
132 if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
133 c->float_to_int16_interleave = float_to_int16_interleave_3dnowext;
136 if (EXTERNAL_SSE(mm_flags)) {
137 c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse;
138 c->float_to_int16 = ff_float_to_int16_sse;
139 c->float_to_int16_interleave = float_to_int16_interleave_sse;
140 c->float_interleave = float_interleave_sse;
142 if (EXTERNAL_SSE2(mm_flags)) {
143 c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse2;
144 c->float_to_int16 = ff_float_to_int16_sse2;
145 c->float_to_int16_interleave = float_to_int16_interleave_sse2;
148 #endif /* HAVE_YASM */