[FFmpeg-devel] [PATCH] swscale: Break loop-carried dependency enabling parallel out of order execution of the gathers.

[FFmpeg-devel] [PATCH] swscale: Break loop-carried dependency enabling parallel out of order execution of the gathers.

[Alan Kelly via ffmpeg-devel]

The gather is unmasked but the instruction does a merge into ymm4, which
depends on the value of ymm4 from the previous loop iteration. The
out-of-order scheduler does not know statically that the instruction is
fully unmasked, preventing parallel out-of-order execution of the
gathers.
---
 libswscale/x86/scale_avx2.asm | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/libswscale/x86/scale_avx2.asm b/libswscale/x86/scale_avx2.asm
index b4b852d60b..90ee8b0a0e 100644
--- a/libswscale/x86/scale_avx2.asm
+++ b/libswscale/x86/scale_avx2.asm
@@ -68,8 +68,10 @@ cglobal hscale8to15_%1, 7, 9, 16, pos0, dst, w, srcmem, filter, fltpos, fltsize,
 .innerloop:
 %endif
     vpcmpeqd  m13, m13
+    pxor m3, m3  ; break loop-carried dependency
     vpgatherdd m3,[srcmemq + m1], m13
     vpcmpeqd  m13, m13
+    pxor m4, m4  ; break loop-carried dependency
     vpgatherdd m4,[srcmemq + m2], m13
     vpunpcklbw m5, m3, m0
     vpunpckhbw m6, m3, m0
@@ -119,6 +121,7 @@ cglobal hscale8to15_%1, 7, 9, 16, pos0, dst, w, srcmem, filter, fltpos, fltsize,
 .tail_innerloop:
 %endif
     vpcmpeqd  xm13, xm13
+    pxor m3, m3  ; break loop-carried dependency
     vpgatherdd xm3,[srcmemq + xm1], xm13
     vpunpcklbw xm5, xm3, xm0
     vpunpckhbw xm6, xm3, xm0
-- 
2.50.1.565.gc32cd1483b-goog

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] swscale: Break loop-carried dependency enabling parallel out of order execution of the gathers.

[Jacob Lifshay]

On August 4, 2025 6:49:20 AM PDT, Alan Kelly via ffmpeg-devel <ffmpeg-devel@ffmpeg.org> wrote:
> The gather is unmasked but the instruction does a merge into ymm4, which
> depends on the value of ymm4 from the previous loop iteration. The
> out-of-order scheduler does not know statically that the instruction is
> fully unmasked, preventing parallel out-of-order execution of the
> gathers.
> ---
>  libswscale/x86/scale_avx2.asm | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/libswscale/x86/scale_avx2.asm b/libswscale/x86/scale_avx2.asm
> index b4b852d60b..90ee8b0a0e 100644
> --- a/libswscale/x86/scale_avx2.asm
> +++ b/libswscale/x86/scale_avx2.asm
> @@ -68,8 +68,10 @@ cglobal hscale8to15_%1, 7, 9, 16, pos0, dst, w, srcmem, filter, fltpos, fltsize,
>  .innerloop:
>  %endif
>      vpcmpeqd  m13, m13
> +    pxor m3, m3  ; break loop-carried dependency

this is in AVX2 code, so you should use vpxor since pxor will just clear the lower 128 bits and leave the upper 128 bits unmodified. actually, on some older intel cpus it will cause a huge stall due to not being v-prefixed:
https://stackoverflow.com/questions/41303780/why-is-this-sse-code-6-times-slower-without-vzeroupper-on-skylake/41349852#41349852

Jacob
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] swscale: Break loop-carried dependency enabling parallel out of order execution of the gathers.

[Hendrik Leppkes]

On Mon, Aug 4, 2025 at 7:19 PM Jacob Lifshay <programmerjake@gmail.com> wrote:
>
>
>
> On August 4, 2025 6:49:20 AM PDT, Alan Kelly via ffmpeg-devel <ffmpeg-devel@ffmpeg.org> wrote:
> > The gather is unmasked but the instruction does a merge into ymm4, which
> > depends on the value of ymm4 from the previous loop iteration. The
> > out-of-order scheduler does not know statically that the instruction is
> > fully unmasked, preventing parallel out-of-order execution of the
> > gathers.
> > ---
> >  libswscale/x86/scale_avx2.asm | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/libswscale/x86/scale_avx2.asm b/libswscale/x86/scale_avx2.asm
> > index b4b852d60b..90ee8b0a0e 100644
> > --- a/libswscale/x86/scale_avx2.asm
> > +++ b/libswscale/x86/scale_avx2.asm
> > @@ -68,8 +68,10 @@ cglobal hscale8to15_%1, 7, 9, 16, pos0, dst, w, srcmem, filter, fltpos, fltsize,
> >  .innerloop:
> >  %endif
> >      vpcmpeqd  m13, m13
> > +    pxor m3, m3  ; break loop-carried dependency
>
> this is in AVX2 code, so you should use vpxor since pxor will just clear the lower 128 bits and leave the upper 128 bits unmodified. actually, on some older intel cpus it will cause a huge stall due to not being v-prefixed:
> https://stackoverflow.com/questions/41303780/why-is-this-sse-code-6-times-slower-without-vzeroupper-on-skylake/41349852#41349852
>

The v is actually automatically added by the pre-processor through
x86inc.asm if the function is marked as avx - its a bit confusing
because all other instructions are explicitly using it however, so it
might still be a good idea to be explicit about it.

As for the patch itself, any numbers?

- Hendrik
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".