Re: [FFmpeg-devel] [PATCH 2/2] vf_blend: Use memcpy when opacity is 0

2016-02-14 Thread Paul B Mahol
On 2/14/16, Timothy Gu wrote: > For xor mode, 1.8x faster on Haswell. > --- > libavfilter/vf_blend.c | 6 ++ > 1 file changed, 6 insertions(+) > > diff --git a/libavfilter/vf_blend.c b/libavfilter/vf_blend.c > index c24013d..599084f 100644 > --- a/libavfilter/vf_blend.c > +++ b/libavfilter/vf

[FFmpeg-devel] [PATCH 2/2] vf_blend: Use memcpy when opacity is 0

2016-02-13 Thread Timothy Gu
For xor mode, 1.8x faster on Haswell. --- libavfilter/vf_blend.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/libavfilter/vf_blend.c b/libavfilter/vf_blend.c index c24013d..599084f 100644 --- a/libavfilter/vf_blend.c +++ b/libavfilter/vf_blend.c @@ -130,6 +130,7 @@ static void blend_c