On Wed, Aug 28, 2019 at 08:36:14PM +0200, Carl Eugen Hoyos wrote:
> Am Mi., 28. Aug. 2019 um 17:46 Uhr schrieb <lance.lmw...@gmail.com>:
> >
> > From: Limin Wang <lance.lmw...@gmail.com>
> >
> > Signed-off-by: Limin Wang <lance.lmw...@gmail.com>
> > ---
> > libavcodec/v210enc.c | 7 +------
> > 1 file changed, 1 insertion(+), 6 deletions(-)
> >
> > diff --git a/libavcodec/v210enc.c b/libavcodec/v210enc.c
> > index 1b840b2..69a2efe 100644
> > --- a/libavcodec/v210enc.c
> > +++ b/libavcodec/v210enc.c
> > @@ -43,12 +43,7 @@ static void v210_planar_pack_8_c(const uint8_t *y, const
> > uint8_t *u,
> > uint32_t val;
> > int i;
> >
> > - /* unroll this to match the assembly */
> > - for (i = 0; i < width - 11; i += 12) {
> > - WRITE_PIXELS(u, y, v, 8);
> > - WRITE_PIXELS(y, u, y, 8);
> > - WRITE_PIXELS(v, y, u, 8);
> > - WRITE_PIXELS(y, v, y, 8);
> > + for (i = 0; i < width - 5; i += 6) {
> > WRITE_PIXELS(u, y, v, 8);
> > WRITE_PIXELS(y, u, y, 8);
> > WRITE_PIXELS(v, y, u, 8);
>
> Doesn't this have a performance impact?
I think it's compiler job for the optimization.
However I have done quick benchmark with ff_v210enc_init_x86 is
removed(to trigger the c function), below is the testing result:
./ffmpeg -benchmark -y -f lavfi -i smptehdbars -sws_flags
+accurate_rnd+bitexact -vf \
scale=720x576,format=yuv422p -c:v v210 -frames 1000 -f framecrc -
master:
frame= 1000 fps=754 q=-0.0 Lsize= 57kB time=00:00:40.00 bitrate=
11.6kbits/s speed=30.2x
bench: utime=1.299s stime=0.013s rtime=1.326s
bench: maxrss=16244736kB
patch applied:
frame= 1000 fps=756 q=-0.0 Lsize= 57kB time=00:00:40.00 bitrate=
11.6kbits/s speed=30.2x
bench: utime=1.300s stime=0.013s rtime=1.323s
bench: maxrss=16166912kB
Thanks,
Limin
>
> Carl Eugen
> _______________________________________________
> 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 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".
