xaymar/FFmpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
101,926 Commits 35 Branches 386 Tags
1050f94c229d4b157467f94de367f87499706c0d
Commit Graph

2467 Commits

Author SHA1 Message Date
Paul B Mahol b69c91bbee avcodec/x86: add cfhdenc SIMD 2021-02-27 17:09:44 +01:00
James Almer f1a894f9d3 avcodec: add missing FF_API_OLD_ENCDEC wrappers to xmm clobber functions 
Signed-off-by: James Almer <jamrial@gmail.com>
 2021-02-26 19:26:31 -03:00
Andreas Rheinhardt 585b764f95 avcodec/x86/constants: Remove unused ff_pw_17 
Unused since 80944df720.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-24 09:49:03 +01:00
Andreas Rheinhardt 7825cc392a avcodec/x86/diracdsp_init: Reuse macro 
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-24 09:38:12 +01:00
Andreas Rheinhardt 0f317eb8e7 avcodec/x86/diracdsp_init: Simplify macro 
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-24 09:36:13 +01:00
Andreas Rheinhardt 68bd6c7dac avcodec/x86/diracdsp_init: Make functions only used here static 
This allowed to remove forward declarations. Because compilers expect
declarations for all functions they encounter even when it is within
blocks disabled via "if (0 && foo)", one has to use a real #if in
ff_diracdsp_init_x86.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-24 09:17:40 +01:00
Andreas Rheinhardt 3a80b1ac12 avcodec/x86/diracdsp_init: Remove unused MMX functions 
Unused since a1f3b18bf5, yet as nonstatic
functions the compiler can't detect this, so that these functions aren't
stripped and no warning is emitted.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-24 08:58:57 +01:00
Andreas Rheinhardt 4f3d8cb554 avcodec/cabac_functions, x86/cabac: Include stddef.h 
Fixes checkheaders after 8c01eb0a31.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2021-02-04 05:17:33 +01:00
Lynne 9e05421dbe ac3enc_fixed: drop unnecessary fixed-point DSP code 2021-01-14 01:44:20 +01:00
Anton Khirnov e15371061d lavu/mem: move the DECLARE_ALIGNED macro family to mem_internal on next+1 bump 
They are not properly namespaced and not intended for public use.
 2021-01-01 14:14:57 +01:00
Anton Khirnov c8c2dfbc37 lavu: move LOCAL_ALIGNED from internal.h to mem_internal.h 
That is a more appropriate place for it.
 2021-01-01 14:11:01 +01:00
Andreas Rheinhardt ead3134150 avcodec/mpegaudiodsp: Make ff_mpadsp_init() thread-safe 
The only thing missing for this is to make ff_mpadsp_init_x86()
thread-safe; it currently isn't because a static table is initialized
every time ff_mpadsp_init() is called (when ARCH_X86 is true). Solve
this by initializing this table only once, namely together with the
ordinary not-arch specific tables. This also allows to reuse their AVOnce.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2020-11-24 11:35:03 +01:00
James Almer 1a35fffaf2 x86/cfhddsp: zero extend int arguments 
if taken from stack, they may have garbage in the upper bits otherwise.
Also, there are only 8 arguments, so don't attempt to load 11.

Fixes SIGSEV crashes in some targets.

Reviewed-by: durandal_1707
Signed-off-by: James Almer <jamrial@gmail.com>
 2020-08-28 20:09:25 -03:00
Paul B Mahol 4aac742505 avcodec/x86/cfhddsp: try to fix build on x32 2020-08-26 23:39:58 +02:00
Paul B Mahol 389cc142fb avcodec/cfhd: add x86 SIMD 
Overall speed changes for 1920x1080, yuv422p10le, 60fps from: 0.19x to 0.343x
 2020-08-26 21:13:38 +02:00
James Almer 2c844c9828 x86/h264_deblock: fix warning about trailing empty parameter 
Fixes part of ticket #8771

Signed-off-by: James Almer <jamrial@gmail.com>
 2020-07-12 11:30:23 -03:00
Martin Storsjö 353aecbb28 pixblockdsp, avdct: Add get_pixels_unaligned 
Use this in vf_spp.c, where the get_pixels operation is done on
unaligned source addresses.

Hook up the x86 (mmx and sse) versions of get_pixels to this
function pointer, as those implementations seem to support unaligned
use.

This fixes fate-filter-spp on armv7.

Signed-off-by: Martin Storsjö <martin@martin.st>
 2020-05-13 13:20:08 +03:00
Linjie Fu 8b8492452d lavc/x86/hevc_add_res: Fix coeff overflow in ADD_RES_SSE_16_32_8 
Fix overflow for coeff -32768 in function ADD_RES_SSE_16_32_8 with no
performance drop.(SSE2/AVX/AVX2)

./checkasm --test=hevc_add_res --bench

Mainline:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_32x32_8_sse2: 127.5
    hevc_add_res_32x32_8_avx: 127.0
    hevc_add_res_32x32_8_avx2: 86.5

Add overflow test case:
  - hevc_add_res.add_residual [FAILED]

After:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_32x32_8_sse2: 126.8
    hevc_add_res_32x32_8_avx: 128.3
    hevc_add_res_32x32_8_avx2: 86.8

Signed-off-by: Xu Guangxin <guangxin.xu@intel.com>
Signed-off-by: Linjie Fu <linjie.fu@intel.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
 2020-03-27 10:57:40 +01:00
Linjie Fu e9abef437f lavc/x86/hevc_add_res: Fix overflow in ADD_RES_SSE_8_8 
Fix overflow for coeff -32768 in function ADD_RES_SSE_8_8 with
no performance drop.

./checkasm --test=hevc_add_res --bench

Mainline:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_8x8_8_sse2: 15.5

Add overflow test case:
  - hevc_add_res.add_residual [FAILED]

After:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_8x8_8_sse2: 15.5

Signed-off-by: Xu Guangxin <guangxin.xu@intel.com>
Signed-off-by: Linjie Fu <linjie.fu@intel.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
 2020-03-27 10:57:40 +01:00
Linjie Fu 0da14ed09e lavc/x86/hevc_add_res: Fix overflow in ADD_RES_MMX_4_8 
Fix overflow for coeff -32768 in function ADD_RES_MMX_4_8 with no
performance drop.

./checkasm --test=hevc_add_res --bench

Mainline:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_4x4_8_mmxext: 15.5

Add overflow test case:
  - hevc_add_res.add_residual [FAILED]

After:
  - hevc_add_res.add_residual [OK]
    hevc_add_res_4x4_8_mmxext: 15.0

Signed-off-by: Xu Guangxin <guangxin.xu@intel.com>
Signed-off-by: Linjie Fu <linjie.fu@intel.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
 2020-03-27 10:57:40 +01:00
Michael Niedermayer 24af459d1e avcodec/x86/diracdsp: Fix high bits on Windows x86_64 
Found-by: james
 2020-01-31 00:04:22 +01:00
Michael Niedermayer 0694b60b7b avcodec/x86/diracdsp: Fix incorrect src addressing in dequant_subband_32() 
Fixes: Segfault (not reproducable with asm, which made this hard to debug)
Fixes: decoding errors
Fixes: 19854/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_DIRAC_fuzzer-5729372837511168

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-01-30 18:47:21 +01:00
Peter Ross fd17218558 vp4: prevent unaligned memory access in loop filter 
VP4 applies a loop filter during motion compensation, causing the block offset
will often by unaligned. This produces a bus error on some platforms, namely
ARMv7 NEON.

This patch adds a unaligned version of the loop filter function pointer
to VP3DSPContext.

Reported-by: Mike Melanson <mike@multimedia.cx>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2019-10-30 10:06:38 +01:00
James Almer 1faedb9a11 x85/opusdsp: enable the functions on all FMA3 CPUs 
It's not using ymm registers, so limiting it to CPUs with fast AVX
is not necessary.

Signed-off-by: James Almer <jamrial@gmail.com>
 2019-09-11 20:50:45 -03:00
James Almer 80444e23ac x86/opusdps: clear the high bits from some gprs 
Fixes checkasm on systems like win64.

Reviewed-by: Lynne
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-09-11 20:42:31 -03:00
James Almer 58d167bcd5 avcodec/Makefile: add missing pngdsp dependency to the lscr decoder 
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-05-14 16:47:56 -03:00
James Almer b41d8ab2e6 x86/v210dec: use named registers 
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-05-03 01:20:18 -03:00
James Almer abf1aa87ab x86/v210dec: don't reserve more xmm regs than needed 
Prevents pointless register saving on win64 for the sse3 and avx
versions of the function.

Signed-off-by: James Almer <jamrial@gmail.com>
 2019-05-03 01:09:50 -03:00
James Almer b0e29357ba x86/v210dec: remove duplicate load instruction 
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-05-03 01:08:34 -03:00
James Darnley 46f1718cd9 avcodec/x86/v210: fix operands of vpblendd used in new avx2 code 
Assembly failed when using yasm rather than nasm.
 2019-05-02 21:20:54 +02:00
Michael Stoner ebd6fb23c5 libavcodec Adding ff_v210_planar_unpack AVX2 
Replaced VSHUFPS with VPBLENDD to relieve port 5 bottleneck
AVX2 is 1.4x faster than AVX
 2019-05-02 19:21:37 +02:00
Lynne 4b7166c9d5 x86/opusdsp: replace loads with shuffles 
Has a slight speedup.
Can't be carried over to aarch64, since it has no shufps-like instruction.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-04-26 20:39:38 -03:00
Lynne b43b8d337d x86/opusdsp: fix WIN64 return value 
Signed-off-by: James Almer <jamrial@gmail.com>
 2019-04-01 11:06:34 -03:00
Lynne 605e330310 x86/opusdsp: implement FMA3 accelerated postfilter and deemphasis 
58893 decicycles in deemphasis_c,  130548 runs,    524 skips
9475 decicycles in deemphasis_fma3,  130686 runs,    386 skips -> 6.21x speedup

24866 decicycles in postfilter_c,   65386 runs,    150 skips
5268 decicycles in postfilter_fma3,   65505 runs,     31 skips -> 4.72x speedup

Total decoder speedup: ~14%

Deemphasis SIMD based on the following unrolling:
const float c1 = CELT_EMPH_COEFF, c2 = c1*c1, c3 = c2*c1, c4 = c3*c1;
float state = coeff;

for (int i = 0; i < len; i += 4) {
    y[0] = x[0] + c1*state;
    y[1] = x[1] + c2*state + c1*x[0];
    y[2] = x[2] + c3*state + c1*x[1] + c2*x[0];
    y[3] = x[3] + c4*state + c1*x[2] + c2*x[1] + c3*x[0];

    state = y[3];
    y += 4;
    x += 4;
}
 2019-04-01 00:22:00 +02:00
Lynne 5468c1d075 celt_pvq_init: only build when CONFIG_OPUS_ENCODER is enabled 
The entire function was defined away before.
 2019-03-31 23:36:43 +02:00
Lynne 4a2c651620 x86/opus_dsp: rename to celt_pvq 
Its only used in the encoder and in CELT's PVQ.
 2019-03-31 23:35:00 +02:00
James Almer d5d699ab6e avcodec/h264dsp: change loop filter stride argument to ptrdiff_t 2019-02-20 15:27:43 -03:00
Martin Vignali 9a22e6fa1d avcodec/proresdsp indent after prev commit 2018-12-02 12:55:35 +01:00
Martin Vignali c097a32e93 avcodec/proresdec : rename dsp part for 10b and check dspinit for supported bits per raw sample 
based on patch by Kieran Kunhya
 2018-12-02 12:55:31 +01:00
Rostislav Pehlivanov 29eb1c51d7 mdct15: simplify x86 exptab permutation 
Removes an unneeded copy and does the 5-point permute in-place.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
 2018-05-07 23:44:40 +01:00
Rostislav Pehlivanov a72d0fb973 mdct15: simplify the fft15 x86 SIMD 
Saves 1 gpr and 2 instructions and simplifies the macros a bit.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
 2018-05-07 23:27:41 +01:00
Kieran Kunhya f9d3841ae6 mpeg4video: Add support for MPEG-4 Simple Studio Profile. 
This is a profile supporting > 8-bit video and has a higher quality DCT
 2018-04-02 13:06:23 +01:00
Aurelien Jacobs f1e490b1ad sbcenc: add MMX optimizations 
This was originally based on libsbc, and was fully integrated into ffmpeg.

Rough speed test:
C version:    speed= 592x
MMX version:  speed= 785x
 2018-03-07 22:26:53 +01:00
Rostislav Pehlivanov 50945482a7 h264_idct: enable unmacro on newer NASM versions 
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
 2018-02-12 10:50:37 +00:00
Martin Vignali 8f9c38b196 avcodec/utvideoenc : add SIMD (avx) for sub_left_prediction 
asm code by Henrik Gramner
 2018-01-28 20:23:11 +01:00
James Almer 6e80079a28 avcodec: increase AV_INPUT_BUFFER_PADDING_SIZE to 64 
AVX-512 support has been introduced, and even if no functions currently
use zmm registers (able to load as much as 64 bytes of consecutive data
per instruction), they will be added eventually.

Reviewed-by: Rostislav Pehlivanov <atomnuker@gmail.com>
Tested-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
 2018-01-11 23:46:31 -03:00
James Almer 438f884fc4 x86/lossless_videodsp: rename ff_add_left_pred_int16_sse4 to ff_add_left_pred_int16_unaligned_ssse3 
SSSE3_FAST is the proper check for it.

Signed-off-by: James Almer <jamrial@gmail.com>
 2017-12-10 00:51:01 -03:00
James Almer a4fc63c0f9 x86/lossless_videodsp: don't overread the dst buffer in ff_add_left_pred_unaligned_avx2 
Fixes valgrind

Signed-off-by: James Almer <jamrial@gmail.com>
 2017-12-10 00:38:05 -03:00
Martin Vignali 630967ef63 avcodec/utvideodec : add SIMD (SSSE3 and AVX2) for gradient_pred 2017-12-09 15:19:03 +01:00
Martin Vignali 4353c35067 avcodec/x86/lossless_videodsp : add avx2 version for add_left_pred 2017-12-09 15:16:03 +01:00