Stowage/ffmpeg
2
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2026-02-17 13:00:24 +00:00
Code Issues Projects Releases Packages Wiki Activity
ffmpeg/libavcodec/x86
History
Rostislav Pehlivanov 70eb77b34e mdct15: add inverse transform postrotation SIMD 
2.5ms frames:
Before   (c):  2638 decicycles in postrotate, 2097040 runs,    112 skips
After (sse3):  1467 decicycles in postrotate, 2097083 runs,     69 skips
After (avx2):  1244 decicycles in postrotate, 2097085 runs,     67 skips

5ms frames:
Before   (c):  4987 decicycles in postrotate, 1048371 runs,    205 skips
After (sse3):  2644 decicycles in postrotate, 1048509 runs,     67 skips
After (avx2):  2031 decicycles in postrotate, 1048523 runs,     53 skips

10ms frames:
Before   (c):  9153 decicycles in postrotate,  523575 runs,    713 skips
After (sse3):  5110 decicycles in postrotate,  523726 runs,    562 skips
After (avx2):  3738 decicycles in postrotate,  524223 runs,     65 skips

20ms frames:
Before   (c): 17857 decicycles in postrotate,  261866 runs,    278 skips
After (sse3): 10041 decicycles in postrotate,  261746 runs,    398 skips
After (avx2):  7050 decicycles in postrotate,  262116 runs,     28 skips

Improves total decoding performance for real world content by 9% with avx2.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
2017-07-30 07:38:39 +01:00
..
aacencdsp.asm aacenc: add SIMD optimizations for abs_pow34 and quantization 2016-10-18 21:41:18 +01:00
aacencdsp_init.c aacenc: add SIMD optimizations for abs_pow34 and quantization 2016-10-18 21:41:18 +01:00
aacpsdsp.asm lavc/aacpsdsp: use ptrdiff_t for stride in hybrid_analysis 2017-06-28 12:22:39 +02:00
aacpsdsp_init.c lavc/aacpsdsp: use ptrdiff_t for stride in hybrid_analysis 2017-06-28 12:22:39 +02:00
ac3dsp.asm
ac3dsp_downmix.asm Merge commit 'b57e38f52c' 2017-03-22 12:49:29 +01:00
ac3dsp_init.c Merge commit 'b57e38f52c' 2017-03-22 12:49:29 +01:00
alacdsp.asm
alacdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
audiodsp.asm Merge commit '6be7944ee2' 2017-03-23 18:05:27 -03:00
audiodsp_init.c Merge commit '12004a9a7f' 2017-03-20 22:35:07 +01:00
blockdsp.asm
blockdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
bswapdsp.asm
bswapdsp_init.c
cabac.h asm: FF_-prefix internal macros used in inline assembly 2016-06-27 17:21:18 +02:00
cavsdsp.c avcodec/x86/cavsdsp: Delete #include "libavcodec/x86/idctdsp.h". 2017-07-21 02:08:33 +02:00
cavsidct.asm cavs: add a sse2 idct implementation. 2017-04-06 10:03:28 -04:00
constants.c vp9mc/x86: add AVX and AVX2 MC 2016-08-03 11:00:08 +02:00
constants.h Merge commit '3c504bc359' 2017-03-15 22:07:28 -03:00
dcadsp.asm x86: Add missing colons after assembly labels 2016-10-17 16:31:26 +02:00
dcadsp_init.c
dct32.asm
dct_init.c
dirac_dwt.asm
dirac_dwt_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
diracdsp.asm x86/diracdsp: make ff_put_signed_rect_clamped_10_sse4 work on x86_32 2016-07-20 13:43:38 -03:00
diracdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
dnxhdenc.asm
dnxhdenc_init.c
fdct.c
fdct.h
fdctdsp_init.c
fft.asm
fft.h
fft_init.c
flac_dsp_gpl.asm
flacdsp.asm
flacdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
fmtconvert.asm
fmtconvert_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
fpel.asm Merge commit '009adfd4fb' 2017-03-21 15:02:31 +01:00
fpel.h
g722dsp.asm
g722dsp_init.c
h263_loopfilter.asm
h263dsp_init.c
h264_cabac.c Merge commit '0a35f128f3' 2017-04-08 14:30:13 +02:00
h264_chromamc.asm Merge commit 'e4a94d8b36' 2017-03-21 15:20:45 -03:00
h264_chromamc_10bit.asm Merge commit 'e4a94d8b36' 2017-03-21 15:20:45 -03:00
h264_deblock.asm avcodec/h264: enable sse2 chroma deblock/loop filter functions 2017-02-27 13:22:06 +01:00
h264_deblock_10bit.asm avcodec/x86: deduplicate PASS8ROWS macro 2017-02-18 20:26:49 +01:00
h264_idct.asm avcodec/h264: add sse2 versions of previous idct functions 2017-05-15 15:00:20 +02:00
h264_idct_10bit.asm avcodec/h264: sse2 and avx 4:2:2 idct add8 10-bit functions 2016-11-30 22:58:28 +01:00
h264_intrapred.asm Merge commit '5801f9ed24' 2017-03-23 11:58:01 +01:00
h264_intrapred_10bit.asm Merge commit '5801f9ed24' 2017-03-23 11:58:01 +01:00
h264_intrapred_init.c h264pred: added AVX2 implementation for tm_vp8 16x16. 2017-03-20 09:45:42 -04:00
h264_qpel.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
h264_qpel_8bit.asm
h264_qpel_10bit.asm x86: Add missing colons after assembly labels 2016-10-17 16:31:26 +02:00
h264_weight.asm x86: Add missing colons after assembly labels 2016-10-17 16:31:26 +02:00
h264_weight_10bit.asm
h264chroma_init.c Merge commit 'e4a94d8b36' 2017-03-21 15:20:45 -03:00
h264dsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
hevc_add_res.asm x86/hevc_add_res: merge last remaining changes from 3d65359832 2017-03-31 20:49:45 -03:00
hevc_deblock.asm avcodec/x86: deduplicate PASS8ROWS macro 2017-02-18 20:26:49 +01:00
hevc_idct.asm Merge commit '112cee0241' 2017-03-23 15:58:46 +01:00
hevc_mc.asm x86: Add missing colons after assembly labels 2016-10-17 16:31:26 +02:00
hevc_sao.asm
hevc_sao_10bit.asm
hevcdsp.h Merge commit '6d5636ad9a' 2017-03-24 12:33:25 +01:00
hevcdsp_init.c Merge commit '6d5636ad9a' 2017-03-24 12:33:25 +01:00
hpeldsp.asm Merge commit '1dfc3cf89d' 2017-01-31 14:49:29 -03:00
hpeldsp.h Revert "Merge commit '0a39c9ac0b'" 2017-02-01 02:01:07 +01:00
hpeldsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
hpeldsp_rnd_template.c asm: FF_-prefix internal macros used in inline assembly 2016-06-27 17:21:18 +02:00
hpeldsp_vp3.asm Merge commit '1dfc3cf89d' 2017-01-31 14:49:29 -03:00
hpeldsp_vp3_init.c Revert "Merge commit '0a39c9ac0b'" 2017-02-01 02:01:07 +01:00
huffyuvdsp.asm huffyuvdsp: move functions only used by huffyuv from lossless_videodsp 2017-01-12 22:53:05 -03:00
huffyuvdsp_init.c lavc/huffyuvdsp: only transmit the pix_fmt instead of the whole avctx 2017-03-22 16:22:20 +01:00
huffyuvencdsp.asm huffyuvencdsp: move shared functions to a new lossless_videoencdsp context 2017-01-12 22:53:04 -03:00
huffyuvencdsp_init.c huffyuvencdsp: move shared functions to a new lossless_videoencdsp context 2017-01-12 22:53:04 -03:00
idctdsp.asm
idctdsp.h avcodec/x86/idctdsp: Remove duplicate include 2017-03-26 19:17:30 +02:00
idctdsp_init.c avcodec/x86: use new x86-64 functions for -idct simple 2017-06-28 17:27:35 +02:00
imdct36.asm avcodec/x86/imdct36: fix building with nasm 2.11.05 2017-01-02 20:44:16 +01:00
inline_asm.h Merge commit '41ed7ab45f' 2016-06-21 21:55:34 +02:00
jpeg2000dsp.asm
jpeg2000dsp_init.c
lossless_audiodsp.asm
lossless_audiodsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
lossless_videodsp.asm Merge commit 'd06dfaa5cb' 2017-01-31 15:36:49 -03:00
lossless_videodsp_init.c avcodec/lossless_videodsp: use ptrdiff_t for length parameters 2017-03-22 18:38:35 -03:00
lossless_videoencdsp.asm huffyuvencdsp: move shared functions to a new lossless_videoencdsp context 2017-01-12 22:53:04 -03:00
lossless_videoencdsp_init.c huffyuvencdsp: move shared functions to a new lossless_videoencdsp context 2017-01-12 22:53:04 -03:00
lpc.c Merge commit '4efab89332' 2017-01-31 15:08:19 -03:00
Makefile avcodec/utvideodec: add SIMD for restore_rgb_planes 2017-06-27 09:54:10 +02:00
mathops.h
mdct15.asm mdct15: add inverse transform postrotation SIMD 2017-07-30 07:38:39 +01:00
mdct15_init.c mdct15: add inverse transform postrotation SIMD 2017-07-30 07:38:39 +01:00
me_cmp.asm
me_cmp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
mlpdsp.asm
mlpdsp_init.c
mpegaudiodsp.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
mpegvideo.c avcodec/x86/mpegvideo: Use intra scantable in dct_unquantize_h263_intra_mmx() 2017-06-20 00:07:51 +02:00
mpegvideodsp.c Merge commit 'dc40a70c57' 2016-06-26 15:53:00 +02:00
mpegvideoenc.c lavc/mpegvideoenc: reformat inv_zigzag_direct16 so the zigzag pattern is visible 2017-05-19 11:17:58 +02:00
mpegvideoenc_qns_template.c
mpegvideoenc_template.c avcodec/x86/mpegenc: support transpose permuation type 2017-06-20 12:12:13 +02:00
mpegvideoencdsp.asm
mpegvideoencdsp_init.c
pixblockdsp.asm Merge commit 'de452e5037' 2017-03-20 15:58:32 +01:00
pixblockdsp_init.c Merge commit 'de452e5037' 2017-03-20 15:58:32 +01:00
pngdsp.asm
pngdsp_init.c
proresdsp.asm avcodec/x86: allow future 8-bit simple idct to use slightly different coefficients 2017-06-20 16:12:25 +02:00
proresdsp_init.c Merge commit '3fd22538bc' 2017-03-19 15:30:13 -03:00
qpel.asm
qpeldsp.asm
qpeldsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
rnd_template.c asm: FF_-prefix internal macros used in inline assembly 2016-06-27 17:21:18 +02:00
rv34dsp.asm x86/rv34dsp: add ff_rv34_idct_dc_add_sse2 2017-02-02 17:51:21 -03:00
rv34dsp_init.c x86/rv34dsp: add ff_rv34_idct_dc_add_sse2 2017-02-02 17:51:21 -03:00
rv40dsp.asm Merge commit '41ed7ab45f' 2016-06-21 21:55:34 +02:00
rv40dsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
sbrdsp.asm Revert "x86/sbrdsp: remove unnecessary sign extend instruction in apply_noise_main" 2017-07-05 10:29:15 -03:00
sbrdsp_init.c
simple_idct.asm avcodec/x86: move simple_idct to external assembly 2017-05-30 13:20:42 +02:00
simple_idct.h avcodec/x86: add an 8-bit simple IDCT function based on the x86-64 high depth functions 2017-06-28 17:27:35 +02:00
simple_idct10.asm avcodec/x86: add an 8-bit simple IDCT function based on the x86-64 high depth functions 2017-06-28 17:27:35 +02:00
simple_idct10_template.asm avcodec/x86: allow future 8-bit simple idct to have "DC only hack" 2017-06-28 17:27:35 +02:00
snowdsp.c asm: FF_-prefix internal macros used in inline assembly 2016-06-27 17:21:18 +02:00
svq1enc.asm
svq1enc_init.c
synth_filter.asm
synth_filter_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
takdsp.asm
takdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
ttadsp.asm avcodec/ttadsp: cosmetics 2016-08-06 18:27:01 -03:00
ttadsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
ttaencdsp.asm x86/ttaenc: add ff_ttaenc_filter_process_{ssse3,sse4} 2016-08-02 15:48:04 -03:00
ttaencdsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
utvideodsp.asm x86/utvideodsp: make restore_rgb_planes functions work on x86_32 2017-07-04 22:49:45 -03:00
utvideodsp_init.c x86/utvideodsp: make restore_rgb_planes functions work on x86_32 2017-07-04 22:49:45 -03:00
v210-init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
v210.asm
v210enc.asm x86: Add missing colons after assembly labels 2016-10-17 16:31:26 +02:00
v210enc_init.c
vc1dsp.h
vc1dsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
vc1dsp_loopfilter.asm
vc1dsp_mc.asm Merge commit '2ec9fa5ec6' 2017-03-21 14:29:52 -03:00
vc1dsp_mmx.c idct: Change type of array stride parameters to ptrdiff_t 2016-09-29 14:48:03 +02:00
videodsp.asm Merge commit 'b89804da9b' 2017-03-23 18:35:49 -03:00
videodsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
vorbisdsp.asm x86/vorbisdsp: optimize ff_vorbis_inverse_coupling_sse 2017-06-15 23:20:05 -03:00
vorbisdsp_init.c
vp3dsp.asm Merge commit '6892df9294' 2017-03-19 18:41:26 +01:00
vp3dsp_init.c Merge commit '6892df9294' 2017-03-19 18:41:26 +01:00
vp6dsp.asm Merge commit 'd9d26a3674' 2017-03-19 14:54:25 -03:00
vp6dsp_init.c Merge commit '721d57e608' 2017-03-19 17:15:24 -03:00
vp8dsp.asm Merge commit '802727b538' 2017-03-19 15:18:31 -03:00
vp8dsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
vp8dsp_loopfilter.asm Merge commit '802727b538' 2017-03-19 15:18:31 -03:00
vp9dsp_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
vp9dsp_init.h vp9: re-split the decoder/format/dsp interface header files. 2017-03-28 18:04:26 -04:00
vp9dsp_init_10bpp.c
vp9dsp_init_12bpp.c
vp9dsp_init_16bpp.c avcodec/x86/vp9dsp_init_16bpp: Fix linking to missing ff_vp9_ipred_dr_32x32_16_avx2() on 32bit 2017-06-28 00:31:33 +02:00
vp9dsp_init_16bpp_template.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00
vp9intrapred.asm
vp9intrapred_16bpp.asm avcodec/vp9: add 64-bit ipred_dr_32x32_16 avx2 implementation 2017-06-27 16:10:50 -04:00
vp9itxfm.asm x86/vp9itxfm: add missing AVX2 guards 2016-11-18 17:01:11 -03:00
vp9itxfm_16bpp.asm
vp9itxfm_template.asm
vp9lpf.asm vp9lpf/x86: make filter_16_h work on 32-bit. 2016-10-04 10:54:09 +02:00
vp9lpf_16bpp.asm doc: fix spelling errors 2016-10-21 23:58:47 +02:00
vp9mc.asm Merge commit 'e99ecda550' 2017-03-16 20:25:39 +01:00
vp9mc_16bpp.asm
vp56_arith.h
w64xmmtest.c Merge commit 'de2ae3c1fa' 2017-03-21 14:43:53 +01:00
xvididct.asm
xvididct.h Merge commit '2ec9fa5ec6' 2017-03-21 14:29:52 -03:00
xvididct_init.c build: Generalize yasm/nasm-related variable names 2017-06-21 17:00:29 -03:00