Thanks for the fix, lgtm as well.
On Tue, Mar 15, 2022 at 12:56 AM Paul B Mahol wrote:
> lgtm
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
Awesome, thanks!
On Thu, Mar 10, 2022 at 9:45 PM Paul B Mahol wrote:
>
>
> On Thu, Mar 10, 2022 at 7:41 PM Victoria Zhislina
> wrote:
>
>> Paul and all, do you have any chances to view my patch from Feb,19? I
>> assume I'\ve fixed all you've kindly pointed
t; zimg_filter_graph_build that used to take 30-60% of each frame
> processig
> >> > only if necessary (some parameters changed)
> >> > the performance increase vs original version
> >> > in video downscale and color conversion >4x is seen
> >> >
and color conversion >4x is seen
> >> > on 64 cores Intel Xeon, 3x on i7-6700K (4 cores with HT)
> >> >
> >> > Signed-off-by: Victoria Zhislina
> >> >
> >>
> >> Can't you patch such a feature into the upstream instead?
>
ce optimization >4x
>
> This reads like an advertisement rather than a useful description. It
> should say what the patch does, performance improvement numbers should
> be mentioned in the commit message body.
>
> Quoting Victoria Zhislina (2022-02-21 09:20:55)
> > B
res Intel Xeon, 3x on i7-6700K (4 cores with HT)
Signed-off-by: Victoria Zhislina
---
libavfilter/vf_zscale.c | 413 +++-
1 file changed, 284 insertions(+), 129 deletions(-)
diff --git a/libavfilter/vf_zscale.c b/libavfilter/vf_zscale.c
index 1288c5e
res Intel Xeon, 3x on i7-6700K (4 cores with HT)
Signed-off-by: Victoria Zhislina
---
libavfilter/vf_zscale.c | 417 +++-
1 file changed, 288 insertions(+), 129 deletions(-)
diff --git a/libavfilter/vf_zscale.c b/libavfilter/vf_zscale.c
index 1288c5e
res Intel Xeon, 3x on i7-6700K (4 cores with HT)
Signed-off-by: Victoria Zhislina
---
libavfilter/vf_zscale.c | 787
1 file changed, 475 insertions(+), 312 deletions(-)
diff --git a/libavfilter/vf_zscale.c b/libavfilter/vf_zscale.c
index 1288c5e
y are absolutely necessary. They save 30-60% of
performance in a single threading case!
On Wed, Feb 9, 2022 at 12:38 PM Guillaume POIRIER
wrote:
> Hello Victoria,
>
> On Sun, 6 Feb 2022 at 16:12, Victoria Zhislina wrote:
>
> > I wasn't aware of the cases where width/heig
res Intel Xeon, 3x on i7-6700K (4 cores with HT)
Signed-off-by: Victoria Zhislina
---
libavfilter/vf_zscale.c | 786
1 file changed, 475 insertions(+), 311 deletions(-)
diff --git a/libavfilter/vf_zscale.c b/libavfilter/vf_zscale.c
index 1288c5e
Paul, thanks a lot for reviewing. The answer is - yes, the move saves up to
30% of performance for a single thread execution, so it is necessary.
I wasn't aware of the cases where width/height can change between frames -
never seen them in real life, but right you are, iI will change my code
accor
, 3x on i7-6700K (4 cores with HT)
Signed-off-by: Victoria Zhislina
---
libavfilter/vf_zscale.c | 779 ++--
1 file changed, 433 insertions(+), 346 deletions(-)
diff --git a/libavfilter/vf_zscale.c b/libavfilter/vf_zscale.c
index 1288c5efc1..1a2de1fe21 100644
12 matches
Mail list logo
