[FFmpeg-devel] [PATCH] VC-1/WMV: vc1 overlap filter asm, MMX/SSE2

[Kostya]

On Tue, Oct 20, 2009 at 04:22:44PM +0200, Michael Niedermayer wrote:
> On Tue, Oct 20, 2009 at 02:39:58PM +0300, Kostya wrote:
> > On Tue, Oct 20, 2009 at 10:55:30AM +0200, Michael Niedermayer wrote:
> > > On Tue, Jun 16, 2009 at 04:11:19PM -0700, Jason Garrett-Glaser wrote:
> > > > Since the patch uses 8x4/4x8 transpose macros, the macros are moved
> > > > into x86util.asm for all of the asm files to use (so it doesn't have
> > > > to be duplicated).  Since this requires %including x86util in the
> > > > deblock code, I had to rename SBUTTERFLY to SBUTTERFLY2 to avoid
> > > > namespace collisions.
> > > > 
> > > > x86_64 overlap_h code is not 100% guaranteed to work; it seems to be
> > > > fine (it only differs in calling convention from x86_32), but every
> > > > x86_64 machine I've tested on gives a different md5sum on every run of
> > > > ffmpeg regardless of whether it's patched or not (at least on my
> > > > samples with overlap filter on).  This is likely a separate bug,
> > > > probably related to the hordes of valgrind errors when decoding wmv3.
> > > > 
> > > > Dark Shikari
> > > 
> > > >  Makefile                  |    1 
> > > >  x86/h264_deblock_sse2.asm |  119 ++++++---------------------------
> > > >  x86/vc1_overlap_mmx.asm   |  163 ++++++++++++++++++++++++++++++++++++++++++++++
> > > >  x86/vc1dsp_mmx.c          |   16 ++++
> > > >  x86/x86util.asm           |   72 ++++++++++++++++++++
> > > >  5 files changed, 276 insertions(+), 95 deletions(-)
> > > > c4b92c83410c251daa0f4dab318f222a1bd38b60  vc1_overlap_asm.diff
> > > 
> > > this patch seems to be lacking a review from our vc1 maintainer
> > > 
> > > and just to make it clear iam fine with any moving or renaming of code that
> > > is needed
> > 
> > As you can see, it only adds SIMD optimisation and has nothing to do
> > with the files I maintain.
> 
> i thought you would also want to maintain vc1 optimisations ...

Not for now, at any rate - my MMX/SSE is rather rusty.
But I have nothing against that patch.
 
> [...]
> -- 
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB