generic-library/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
60,430 Commits 24 Branches 241 Tags 173 MiB
3b6655ebff
Commit Graph

1433 Commits

Author SHA1 Message Date
Clément Bœsch
669d4f9053 x86/vp9lpf: simplify 2nd transpose in 44/48/88/84. 
For non-avx optims, this saves 8 movs.

before:
  1785 decicycles in ff_vp9_loop_filter_h_44_16_ssse3, 524129 runs, 159 skips
  3327 decicycles in ff_vp9_loop_filter_h_48_16_ssse3, 262116 runs, 28 skips
  2712 decicycles in ff_vp9_loop_filter_h_88_16_ssse3, 4193729 runs, 575 skips
  3237 decicycles in ff_vp9_loop_filter_h_84_16_ssse3, 524061 runs, 227 skips

after:
  1768 decicycles in ff_vp9_loop_filter_h_44_16_ssse3, 524062 runs, 226 skips
  3310 decicycles in ff_vp9_loop_filter_h_48_16_ssse3, 262107 runs, 37 skips
  2719 decicycles in ff_vp9_loop_filter_h_88_16_ssse3, 4193954 runs, 350 skips
  3184 decicycles in ff_vp9_loop_filter_h_84_16_ssse3, 524236 runs, 52 skips
 2014-02-08 11:10:23 +01:00
Michael Niedermayer
82ae8a44e6 Merge commit '5b59a9fc6152169599561f04b4f66370edda5c9c' 
* commit '5b59a9fc6152169599561f04b4f66370edda5c9c':
  x86: dcadsp: implement int8x8_fmul_int32

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-02-08 01:20:33 +01:00
Christophe Gisquet
5b59a9fc61 x86: dcadsp: implement int8x8_fmul_int32 
For the callable function (as opposed to the inline one):
         C  SSE  SSE2  SSE4
Win32:  47   42   29    26
Win64:  30   33   25    23
The SSE version is neither compiled nor set for ARCH_X86_64, as the
inlinable function takes over.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
 2014-02-07 22:52:40 +01:00
Loren Merritt
9c978f243a flac/x86: add ff_flac_lpc_32_sse4() 
benchmarked on sandybridge x86_64:
1358232 decicycles in flac_lpc_32_c
1244575 decicycles in flac_lpc_32_sse4, James Almer's patch
 650045 decicycles in flac_lpc_32_sse4, this patch

I haven't tested the edgecases such as odd block lengths

odd block length tested-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-02-06 02:51:19 +01:00
Clément Bœsch
d92a725329 x86/vp9lpf: remove 8 SWAPs in 84/48 transpose. 2014-02-05 07:21:13 +01:00
Clément Bœsch
97dde561de x86/vp9lpf: remove braindead double pxor. 2014-02-05 07:21:11 +01:00
Clément Bœsch
9a3b05b0a9 x86/vp9lpf: save a few mov in flat8in/hev masks calc. 2014-02-05 07:21:09 +01:00
Clément Bœsch
91d85bb167 x86/vp9lpf: add ff_vp9_loop_filter_[vh]_44_16_{sse2,ssse3,avx}. 2014-02-05 07:21:06 +01:00
Michael Niedermayer
de17ccc774 Merge commit '51daafb02eaf96e0743a37ce95a7f5d02c1fa3c2' 
* commit '51daafb02eaf96e0743a37ce95a7f5d02c1fa3c2':
  x86: videodsp: Properly mark sse2 instructions in emulated_edge_mc as such.

Conflicts:
	libavcodec/x86/videodsp_init.c

See: 1b3a7e1f42
Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-31 14:30:30 +01:00
Clément Bœsch
c5dd73b890 x86/vp9lpf: add ff_vp9_loop_filter_h_{48,84}_16_{sse2,ssse3,avx}(). 
5.40s → 5.30s overall decode time with -threads 1 on ped1080p.webm
(i7 920, ssse3)
 2014-01-30 19:34:13 +01:00
Ronald S. Bultje
9ee9c679a7 x86: videodsp: Fix a bug in a %if statement where we used '%%' instead of '&&'. 
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
 2014-01-30 15:33:23 +01:00
Ronald S. Bultje
51daafb02e x86: videodsp: Properly mark sse2 instructions in emulated_edge_mc as such. 
Should fix crashes or corrupt output on pre-SSE2 CPUs when they were
using SSE2-code (e.g. AMD Athlon XP 2400+ or Intel Pentium III) in
hfix or hvar single-edge (left/right) extension functions.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
 2014-01-30 15:30:01 +01:00
James Almer
644c32ea4b x86/vp9lpf: add ff_vp9_loop_filter_[vh]_88_16_sse2() 
Similar gains as the ssse3 version once again

Signed-off-by: James Almer <jamrial@gmail.com>
 2014-01-28 09:30:55 +01:00
Clément Bœsch
222c46c531 x86/vp9lpf: add ff_vp9_loop_filter_[vh]_88_16_{ssse3,avx}. 
9680 decicycles in loop_filter_v_88_16_c, 4193765 runs, 539 skips
9233 decicycles in loop_filter_h_88_16_c, 4193751 runs, 553 skips

1929 decicycles in ff_vp9_loop_filter_v_88_16_ssse3, 4194118 runs, 186 skips
2738 decicycles in ff_vp9_loop_filter_h_88_16_ssse3, 4193861 runs, 443 skips

5.978 → 5.417 overall decode time on ped1080p.webm (-threads 1)

Adding SSE2 support should be relatively trivial (just a matter of
changing the pshufb [mask_mix] with something else), patch welcome.
 2014-01-28 07:36:38 +01:00
Clément Bœsch
822385d775 x86/vp9lpf: add a preload system in FILTER_UPDATE. 
Allow some macro refactoring in filter14().
 2014-01-27 22:39:26 +01:00
Clément Bœsch
315b4775ad x86/vp9lpf: refactor v/h using common macros for P7 to Q7. 2014-01-27 22:39:26 +01:00
Clément Bœsch
5d144086cc x86/vp9lpf: faster P7..Q7 accesses. 
Introduce 2 additional registers for stride3 and mstride3 to allow
direct accesses (lea drops).

3931 → 3827 decicycles in ff_vp9_loop_filter_v_16_16_ssse3

Also uses defines to clarify the code.
 2014-01-27 22:37:42 +01:00
Clément Bœsch
5f4d04d084 x86/lossless_videodsp: silly one-line cosmetic. 2014-01-25 16:24:50 +01:00
Clément Bœsch
5267e85056 x86/lossless_videodsp: use common macro for add and diff int16 loop. 2014-01-25 14:27:37 +01:00
Clément Bœsch
cddbfd2a95 x86/lossless_videodsp: simplify and explicit aligned/unaligned flags 2014-01-25 11:59:43 +01:00
Ronald S. Bultje
c9e6325ed9 vp9/x86: use explicit register for relative stack references. 
Before this patch, we explicitly modify rsp, which isn't necessarily
universally acceptable, since the space under the stack pointer might
be modified in things like signal handlers. Therefore, use an explicit
register to hold the stack pointer relative to the bottom of the stack
(i.e. rsp). This will also clear out valgrind errors about the use of
uninitialized data that started occurring after the idct16x16/ssse3
optimizations were first merged.
 2014-01-24 19:25:25 -05:00
Ronald S. Bultje
97474d527f vp9/x86: iwht4x4 (lossless) mmx. 2014-01-24 19:25:25 -05:00
Ronald S. Bultje
d43efa68bd vp9/x86: 4x4 iadst SIMD (ssse3) variants. 
Cycle measurements for intra itxfm_4x4_add on ped1080p.webm:
idct_idct:    66 -> 67 cycles (noise measurement)
idct_iadst:  199 -> 79 cycles
iadst_idct:  165 -> 70 cycles
iadst_iadst: 183 -> 82 cycles
 2014-01-24 19:25:25 -05:00
Ronald S. Bultje
baf47020cd vp9/x86: 8x8 iadst SIMD (ssse3/avx) variants. 
Cycle measurements for intra itxfm_8x8_add on ped1080p.webm:
idct_idct:   133 -> 135 cycles (noise measurement)
idct_iadst:  900 -> 241 cycles
iadst_idct:  864 -> 215 cycles
iadst_iadst: 973 -> 310 cycles
 2014-01-24 19:25:25 -05:00
Michael Niedermayer
e6d1c66d74 avcodec/x86/lossless_videodsp: disable median optimizations for 16bps 
They only support upto 15bps

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-23 01:51:24 +01:00
Michael Niedermayer
eaacfc7dd1 avcodec/lossless_videodsp: Pass AVCodecContext to init 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-23 01:43:00 +01:00
Michael Niedermayer
ef00ef7553 avcodec/x86/lossless_videodsp: port sub_hfyu_median_prediction_int16 to yasm 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-22 23:27:27 +01:00
Michael Niedermayer
fad49aae28 avcodec/x86/lossless_videodsp: Port sub_hfyu_median_prediction_mmxext to int16 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-22 22:55:49 +01:00
Michael Niedermayer
fee97f25fa avcodec/x86/lossless_videodsp: port add_hfyu_median_prediction_mmxext to 16bit 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-22 21:11:40 +01:00
Michael Niedermayer
631939bde6 avcodec/x86/lossless_videodsp: add diff_int16_mmx/sse2 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-22 19:41:21 +01:00
Reimar Döffinger
76421982d0 lossless_videodsp.asm: fix compilation. 
Fixes these errors with nasm:
libavcodec/x86/lossless_videodsp.asm:86: error: invalid combination of opcode and operands
libavcodec/x86/lossless_videodsp.asm:88: error: invalid combination of opcode and operands
I don't know whether movd or movq was meant, but either way
maskq vs. maskd must match the mov size.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
 2014-01-21 19:46:02 +01:00
Michael Niedermayer
83b67ca056 avcodec/x86/lossless_videodsp: Port lorens add_hfyu_left_prediction_ssse3/sse4 to 16bit 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-21 02:55:41 +01:00
Michael Niedermayer
63d2be7533 avcodec/x86/lossless_videodsp: use SPLATW in add_int16 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-21 02:33:20 +01:00
Michael Niedermayer
f70d7eb20c Move add/diff_int16 to lossless_videodsp 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-20 21:32:47 +01:00
Michael Niedermayer
a493f8541d avcodec/x86/dsp: add_int16_mmx / add_int16_sse2 
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-20 04:06:46 +01:00
James Almer
26800e3864 vp9/x86: rename ff_avg[48]_sse to ff_avg[48]_mmxext 
pavgb is an sse integer instruction, so the mmxext flag is enough

Signed-off-by: James Almer <jamrial@gmail.com>
Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-18 17:08:25 +01:00
James Almer
d2a7314f1e vp9/x86: add ff_vp9_loop_filter_[vh]_16_16_sse2(). 
Similar gains in performance as the SSSE3 version

Signed-off-by: James Almer <jamrial@gmail.com>
 2014-01-17 14:16:38 +01:00
Ronald S. Bultje
8173d1ffc0 vp9/x86: 16x16 iadst_idct, idct_iadst and iadst_iadst (ssse3+avx). 
Sample timings on ped1080p.webm (of the ssse3 functions):
iadst_idct:  4672 -> 1175 cycles
idct_iadst:  4736 -> 1263 cycles
iadst_iadst: 4924 -> 1438 cycles
Total decoding time changed from 6.565s to 6.413s.
 2014-01-16 13:49:31 +01:00
Clément Bœsch
9cc8fa63dd vp9/x86: simplify a few mc inits. 2014-01-16 07:48:27 +01:00
Michael Niedermayer
6391dec82a Merge remote-tracking branch 'qatar/master' 
* qatar/master:
  x86: dsputil: Simplify xvmc deprecation conditional

Conflicts:
	libavcodec/x86/dsputil_init.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-15 20:41:08 +01:00
Diego Biurrun
aab40bbfd5 x86: dsputil: Simplify xvmc deprecation conditional 2014-01-15 15:23:46 +01:00
Clément Bœsch
8b4190da93 vp9/x86: add AVX for itxfm and lpf. 
4412 decicycles in ff_vp9_loop_filter_h_16_16_ssse3, 4193462 runs, 842 skips
3600 decicycles in ff_vp9_loop_filter_h_16_16_avx, 4193621 runs, 683 skips

3010 decicycles in ff_vp9_loop_filter_v_16_16_ssse3, 4193528 runs, 776 skips
2678 decicycles in ff_vp9_loop_filter_v_16_16_avx, 4193742 runs, 562 skips

23025 decicycles in ff_vp9_idct_idct_32x32_add_ssse3, 2096871 runs, 281 skips
19943 decicycles in ff_vp9_idct_idct_32x32_add_avx, 2096815 runs, 337 skips

4675 decicycles in ff_vp9_idct_idct_16x16_add_ssse3, 4194018 runs, 286 skips
3980 decicycles in ff_vp9_idct_idct_16x16_add_avx, 4194022 runs, 282 skips

967 decicycles in ff_vp9_idct_idct_8x8_add_ssse3, 16776972 runs, 244 skips
887 decicycles in ff_vp9_idct_idct_8x8_add_avx, 16777002 runs, 214 skips
 2014-01-15 15:54:03 +01:00
Michael Niedermayer
cb613657ee avcodec/x86/proresdsp_init: x86 prores IDCT is bitexact again 
reenable it for for bitexact mode

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-14 15:59:00 +01:00
Michael Niedermayer
b148a39d55 Merge commit '46bacb5cc6169ff5e8e982495c4925467c1d8bb7' 
* commit '46bacb5cc6169ff5e8e982495c4925467c1d8bb7':
  x86: Consistently use cpu flag detection macros in places that still miss it

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-14 14:44:59 +01:00
Diego Biurrun
46bacb5cc6 x86: Consistently use cpu flag detection macros in places that still miss it 2014-01-14 00:04:58 +01:00
Clément Bœsch
af68bd1c06 vp9/x86: add ff_vp9_loop_filter_[vh]_16_16_ssse3(). 
16662 decicycles in loop_filter_h_16_16_c, 8387355 runs, 1253 skips
17510 decicycles in loop_filter_v_16_16_c, 8387516 runs, 1092 skips

4941 decicycles in ff_vp9_loop_filter_h_16_16_ssse3, 8387887 runs, 721 skips
3899 decicycles in ff_vp9_loop_filter_v_16_16_ssse3, 8387980 runs, 628 skips

Overall decode time goes from:
  ./ffmpeg -v 0 -nostats -threads 1 -i ~/samples/vp9/ped1080p.webm -f null -  8.10s user 0.02s system 99% cpu 8.126 total
to:
  ./ffmpeg -v 0 -nostats -threads 1 -i ~/samples/vp9/ped1080p.webm -f null -  6.15s user 0.04s system 99% cpu 6.199 total

(46 to 61 fps)
 2014-01-12 20:20:24 +01:00
Clément Bœsch
e11ceea68f vp9/x86: factor out some code in VP9_UNPACK_MULSUB_2W_4X. 2014-01-12 20:19:00 +01:00
Clément Bœsch
c9aa0b8f70 vp9/x86: remove reg redundancy in VP9_MULSUB_2W_2X. 2014-01-12 20:18:55 +01:00
Clément Bœsch
7c55ee6168 vp9/x86: merge IDCT coef macros. 2014-01-12 20:18:44 +01:00
Michael Niedermayer
92b2404571 Merge commit '4c642d8d98703faf52983243098f35865e15b312' 
* commit '4c642d8d98703faf52983243098f35865e15b312':
  x86: hpeldsp: Add missing av_cold attribute to init function

Merged-by: Michael Niedermayer <michaelni@gmx.at>
 2014-01-09 20:32:53 +01:00