[FFmpeg-devel] [PATCH] Re: Color corruption and seeking errors with H264 disc sources

[Guillaume POIRIER]

Hi,

On 7/5/07, Michael Niedermayer <michaelni at gmx.at> wrote:
> Hi
>
> On Thu, Jul 05, 2007 at 11:06:44AM +0200, Andreas ?man wrote:
> > Hello,
> >
> > Michael Niedermayer wrote:
> > >Hi
> > >
> > >On Tue, Jul 03, 2007 at 01:32:07PM +0200, Andreas ?man wrote:
> > Okay, i've done my homework, split them up and run some more detailed
> > benchmarks.
> >
> > I've then applied qmul.diff and qplookup.diff and run tests
> > on these optimizations.
> >
> > Then i fixed the actual bug in secondqp.diff
> >
> > So, these three patches needs to be applied in that order.
> >
> > try#1 is the same benchmark with the bugfix but without the
> > optimizations (the patch i sent earlier)
> >
> > The diff between the svn and secondqp median is in the 0.1-0.3%
> > range if i'm not mistaken. So i believe it's quite okay.
> >
> > Anyway, this is my last shot at this right now, i'd rather
> > focus on the h264-multithreading performance issues thats
> > left.

[..]


> qmul.diff and qplookup.diff are ok

Andreas, would you care to provide short descriptions for each patch
so that the commit messages can be meaningful?

Thanks a lot!

Guillaume
-- 
Subversion has been the most pointless project ever started
  -- Linus Torvalds