Jason Garrett-Glaser 19fb234e4a H.264: split luma dc idct out and implement MMX/SSE2 versions ...

About 2.5x the speed.

NOTE: the way that the asm code handles large qmuls is a bit suboptimal.
If x264-style dequant was used (separate shift and qmul values), it might
be possible to get some extra speed.

Originally committed as revision 26336 to svn://svn.ffmpeg.org/ffmpeg/trunk

2011-01-14 21:34:25 +00:00

..

cavsdsp_mmx.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

dct32_sse.c

xmm_clobbers: list xmm registers first in clobber list

2010-10-31 18:14:48 +00:00

deinterlace.asm

Use SECTION .text for yasm code.

2010-12-01 13:12:39 +00:00

dnxhd_mmx.c

dnxhd_mmx: prefer xmm registers below xmm6 when they are available

2010-11-02 03:09:16 +00:00

dsputil_mmx_avg_template.c

…

dsputil_mmx_qns_template.c

…

dsputil_mmx_rnd_template.c

…

dsputil_mmx.c

H.264: split luma dc idct out and implement MMX/SSE2 versions

2011-01-14 21:34:25 +00:00

dsputil_mmx.h

Fix compilation on x86-32 with --disable-optimizations,

2011-01-03 11:30:04 +00:00

dsputil_yasm.asm

…

dsputilenc_mmx.c

dsputil_mmx: prefer xmm registers below xmm6 when they are available

2010-10-31 13:13:53 +00:00

dsputilenc_yasm.asm

Don't access upper 32 bits of a 32-bit int on 64-bit systems.

2010-09-17 12:24:22 +00:00

fdct_mmx.c

cosmetics: split long line

2010-10-31 13:46:17 +00:00

fft_3dn2.c

imdct/x86: Use "s->mdct_size" instead of "1 << s->mdct_bits".

2010-08-23 15:51:09 +00:00

fft_3dn.c

…

fft_mmx.asm

Split and then simplify address generation macro.

2010-09-26 09:08:11 +00:00

fft_sse.c

fft: mark xmm registers as clobbered in ff_imdct_calc_sse

2010-10-06 01:27:02 +00:00

fft.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

fft.h

SSE optimized 32-point DCT

2010-07-06 16:58:54 +00:00

h264_chromamc.asm

For rounding in chroma MC SSSE3, use 16-byte pw_3/4 instead of reading 8 bytes

2010-12-24 17:23:22 +00:00

h264_deblock.asm

Port latest x264 deblock asm (before they moved to using NV12 as internal

2010-09-03 16:52:46 +00:00

h264_i386.h

…

h264_idct.asm

H.264: split luma dc idct out and implement MMX/SSE2 versions

2011-01-14 21:34:25 +00:00

h264_intrapred_init.c

Port pred8x8l_down_left_mmxext (H.264 intra prediction) from x264 (authors:

2010-12-29 23:48:44 +00:00

h264_intrapred.asm

Fix invalid reads in valgrind fate, patch by Daniel Kang <daniel dot d dot

2010-12-31 01:29:06 +00:00

h264_qpel_mmx.c

xmm_clobbers: list xmm registers first in clobber list

2010-10-31 18:14:48 +00:00

h264_weight.asm

Use "d" suffix for general-purpose registers used with movd.

2010-09-05 10:10:16 +00:00

h264dsp_mmx.c

H.264: split luma dc idct out and implement MMX/SSE2 versions

2011-01-14 21:34:25 +00:00

idct_mmx_xvid.c

…

idct_mmx.c

Fix compilation in x86_64. I broke it with r24580.

2010-07-29 22:45:21 +00:00

idct_sse2_xvid.c

xmm_clobbers: list xmm registers first in clobber list

2010-10-31 18:14:48 +00:00

idct_xvid.h

…

lpc_mmx.c

lpc_mmx: add xmm registers to clobber list

2010-10-31 23:37:15 +00:00

Makefile

Move sse16_sse2() from inline asm to yasm. It is one of the functions causing

2010-09-17 01:44:17 +00:00

mathops.h

…

mlpdsp.c

…

motion_est_mmx.c

sad16_sse2: merge 2 asm blocks

2010-10-31 21:20:20 +00:00

mpegaudiodec_mmx.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

mpegvideo_mmx_template.c

mpegvideo_mmx: add xmm registers to clobber list

2010-10-31 13:15:16 +00:00

mpegvideo_mmx.c

mpegvideo_mmx: add xmm registers to clobber list

2010-10-31 13:15:16 +00:00

simple_idct_mmx.c

Put ff_ prefix on non-static {put_signed,put,add}_pixels_clamped_mmx()

2010-08-30 16:22:27 +00:00

snowdsp_mmx.c

snowdsp: Explicitly state the operand sizes

2010-10-04 13:08:13 +00:00

vc1dsp_mmx.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

vc1dsp_yasm.asm

Use "d" suffix for general-purpose registers used with movd.

2010-09-05 10:10:16 +00:00

vp3dsp.asm

Use "d" suffix for general-purpose registers used with movd.

2010-09-05 10:10:16 +00:00

vp8dsp-init.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

vp8dsp.asm

Use "d" suffix for general-purpose registers used with movd.

2010-09-05 10:10:16 +00:00

vp56_arith.h

VP5/6/8: ~7% faster arithmetic decoding

2010-08-12 01:11:32 +00:00

vp56dsp_init.c

Move mm_support() from libavcodec to libavutil, make it a public

2010-09-08 15:07:14 +00:00

vp56dsp.asm

Fix typos when converting inline asm to yasm, fixes MMX-only fate-ea-vp61.

2010-08-26 14:33:39 +00:00

x86inc.asm

sync yasm macros from x264

2010-07-21 22:45:16 +00:00

x86util.asm

Move hadamard_diff{,16}_{mmx,mmx2,sse2,ssse3}() from inline asm to yasm,

2010-09-17 01:56:06 +00:00