Performance of XRenderCompositeTrapezoids in the aliased case

[Mart Raudsepp]

On Sat, 2009-02-14 at 19:18 +0100, Michel Dänzer wrote:
> On Fri, 2009-02-06 at 18:25 +0100, Michel Dänzer wrote:
> > On Tue, 2009-02-03 at 04:10 +0200, Mart Raudsepp wrote:
> > > 
> > > I shamefully admit I haven't actually gotten around to pinpoint this
> > > glyph perf regression with 101% certainty at Pict_A8 OpAdd's, but
> > > logically that should be the case.
> > 
> > Does the attached patch help? It should promote a8 masks to argb if the
> > driver can't composite to a8. Beware, I only tested that it compiles and
> > doesn't seem to break anything else.
> 
> Here's a slightly better patch, though there shouldn't be any functional
> difference.
> 
> Did you get a chance to try it?

Sorry not yet, but I'm building up a Geode linux system at home as we
speak to work on that stuff from here.

> BTW, this isn't useful only for Geode but also for other setups that
> can't composite to a8, e.g. i830 class or some older cards supported by
> nouveau I think.

Indeed.


Regards,
Mart Raudsepp