render improvements

[Michel Dänzer]

On Sat, 2005-16-04 at 20:48 -0400, Zack Rusin wrote:
> On Saturday 16 April 2005 18:59, Owen Taylor wrote:
> > On Fri, 2005-04-15 at 20:35 -0400, Zack Rusin wrote:
> > > > The benchmark numbers you posted look good for the general case
> > > > but are marginally slower for Over blends, which will probably be
> > > > 90% of Render usage.  I suspect this is a result of not using the
> > > > MMX path.
> > >
> > > Right. The super-fancy special casing is not in ;) Once Owen
> > > commits the MMX code to xserver we'll merge that back in and I'll
> > > add some AltiVec code because we just can't have PPC trailing
> > > behind, can we? ;)
> >
> > This was committed yesterday. (Had to get Soeren to do it for me
> > since I don't have commit access to the xserver tree, as it turns
> > out.)
> 
> Great, thanks. You forgot to commit the fbmmx.{h,c} files though. I 
> merged in our changes anyway. The new diff is at 
> http://ktown.kde.org/~zrusin/render_xserver.diff . But this one is 
> completely untested (I have to leave for a bit. It compiles, hence we 
> can ship it, we just might not want to be running it ;) ) .

The MMX code in the xserver tree is broken here (only the shadows of
translucent windows are visible), but it seems to be the same with or
without your patch.


-- 
Earthling Michel Dänzer      |     Debian (powerpc), X and DRI developer
Libre software enthusiast    |   http://svcs.affero.net/rm.php?r=daenzer