Re: [FFmpeg-devel] [PATCH v2 2/3] avfilter/x86/vf_exposure: add ff_exposure_avx2

James Almer: >On 11/20/2021 5:42 PM, Wu Jianhua wrote: >> James Almer: >> On 11/4/2021 1:18 AM, Wu Jianhua wrote: Performance data(Less is better): exposure_sse: 500491 >> >>> You reported a better result in the first patch. >> >

[FFmpeg-devel] [PATCH] avcodec/h263: Fix global-buffer-overflow with noout flag2 set

h263_get_motion_length() forgot to take an absolute value; as a consequence, a negative index was used to access an array. This leads to potential crashes, but mostly it just accesses what is to the left of ff_mvtab (unless one uses ASAN), thereby defeating the purpose of the AV_CODEC_FLAG2_NO_OUTP

Re: [FFmpeg-devel] [PATCH 1/2] av(codec|device): Don't cast pointers to int

On Sat, Nov 20, 2021 at 05:32:56PM +0100, Andreas Rheinhardt wrote: > C99/C11 6.3.2.3 5: "Any pointer type may be converted to an integer > type. [...] If the result cannot be represented in the integer type, > the behavior is undefined." So stop casting pointers to int; use > uintptr_t instead. >

Re: [FFmpeg-devel] [PATCH v2 2/3] avfilter/x86/vf_exposure: add ff_exposure_avx2

On 11/20/2021 5:42 PM, Wu Jianhua wrote: James Almer: On 11/4/2021 1:18 AM, Wu Jianhua wrote: Performance data(Less is better): exposure_sse: 500491 You reported a better result in the first patch. For they are tested on different baseline, I think it might

Re: [FFmpeg-devel] [PATCH v2 2/3] avfilter/x86/vf_exposure: add ff_exposure_avx2

James Almer: On 11/4/2021 1:18 AM, Wu Jianhua wrote: >> Performance data(Less is better): >> exposure_sse: 500491 >You reported a better result in the first patch. For they are tested on different baseline, I think it might be better to only compare these two val

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/h263: Inline constant

On Fri, Nov 19, 2021 at 08:16:52PM +0100, Andreas Rheinhardt wrote: > Signed-off-by: Andreas Rheinhardt > --- > libavcodec/h263.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/libavcodec/h263.h b/libavcodec/h263.h > index 491f2e0aac..70fd1ffdc0 100644 > --- a/libavcode

Re: [FFmpeg-devel] [PATCH v2 3/3] tests/checkasm: add check for vf_exposure

On 11/4/2021 1:18 AM, Wu Jianhua wrote: Signed-off-by: Wu Jianhua --- tests/checkasm/Makefile | 1 + tests/checkasm/checkasm.c| 3 ++ tests/checkasm/checkasm.h| 1 + tests/checkasm/vf_exposure.c | 62 tests/fate/checkasm.mak | 1 +

Re: [FFmpeg-devel] [PATCH v2 2/3] avfilter/x86/vf_exposure: add ff_exposure_avx2

On 11/4/2021 1:18 AM, Wu Jianhua wrote: Performance data(Less is better): exposure_sse: 500491 You reported a better result in the first patch. exposure_avx2: 449122 This looks like a really low speed up for a function that processes twice the amount of floats per loop. Si

Re: [FFmpeg-devel] [PATCH v2 1/3] avfilter/x86/vf_exposure: add x86 SIMD optimization

On 11/20/2021 1:46 PM, James Almer wrote: On 11/4/2021 1:18 AM, Wu Jianhua wrote: diff --git a/libavfilter/x86/vf_exposure.asm b/libavfilter/x86/vf_exposure.asm new file mode 100644 index 00..3351c6fb3b --- /dev/null +++ b/libavfilter/x86/vf_exposure.asm @@ -0,0 +1,55 @@ +;*

Re: [FFmpeg-devel] [PATCH v2 1/3] avfilter/x86/vf_exposure: add x86 SIMD optimization

On 11/4/2021 1:18 AM, Wu Jianhua wrote: Performance data(Less is better): exposure_c:857394 exposure_sse: 327589 Signed-off-by: Wu Jianhua --- libavfilter/exposure.h | 36 +++ libavfilter/vf_exposure.c | 36 +-- libavfilter/x86/Makef

[FFmpeg-devel] [PATCH 2/2] Revert "Disable warnings for casting pointers to integers, there is nothing wrong with that."

This reverts commit 5258f64a14713499cf84840b3ab3a1ee7cdcaeb8. The premise of said commit (that conversions from pointer to int are ok) is wrong: C99/C11 6.3.2.3 5: "Any pointer type may be converted to an integer type. [...] If the result cannot be represented in the integer type, the behavior is u

[FFmpeg-devel] [PATCH 1/2] av(codec|device): Don't cast pointers to int

C99/C11 6.3.2.3 5: "Any pointer type may be converted to an integer type. [...] If the result cannot be represented in the integer type, the behavior is undefined." So stop casting pointers to int; use uintptr_t instead. Signed-off-by: Andreas Rheinhardt --- When did these asserts actually help t

[FFmpeg-devel] [PATCH] avformat/cafdec: Avoid unnecessary avio_tell() calls

Signed-off-by: Andreas Rheinhardt --- libavformat/cafdec.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/cafdec.c b/libavformat/cafdec.c index 3ff0558afe..4b5b15b58d 100644 --- a/libavformat/cafdec.c +++ b/libavformat/cafdec.c @@ -224,7 +224,7 @@ static int r

Re: [FFmpeg-devel] [PATCH v2 1/3] avfilter/x86/vf_exposure: add x86 SIMD optimization

will apply soon ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH] fftools/ffmpeg: Take type limitations of AVFifo API into account

The types used by the AVFifo API are inconsistent: av_fifo_(space|size)() returns an int; av_fifo_alloc() takes an unsigned, other parts use size_t. This commit therefore ensures that the size of the muxing_queue FIFO never exceeds INT_MAX. While just at it, also make sure not to call av_fifo_size

Re: [FFmpeg-devel] [PATCH 2/2] libRIST: allow setting fifo size and fail on overflow.

Thanks for reviewing again! On 18-11-2021 20:35, Marton Balint wrote: On Wed, 17 Nov 2021, Gijs Peskens wrote: Introduce fifo_size and overrun_nonfatal params to configure fifo buffer behavior. Fifo size is used to left shift 2, since libRIST only accepts powers of 2. Can you please make

Re: [FFmpeg-devel] [PATCH] avcodec/smcenc: Avoid dangling pointers in context

Paul B Mahol: > Commit message is misleading, better to use: switch to stack for > PutByteContext > Will do. Although the commit message is correct: The pointers contained in the PutByteContext become dangling after each encode call. - Andreas ___ ffmpe

Re: [FFmpeg-devel] [PATCH] avcodec/smcenc: Avoid dangling pointers in context

Commit message is misleading, better to use: switch to stack for PutByteContext ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.