Re: FYI: 15x 2D performance decrease in radeon driver KMS
What would you recommend to mesure / test them?
On Sat, May 29, 2010 at 04:46:25PM +0800, Zhang Jiarui wrote:
> How about the 2D performance of web browser?
> On Sat, May 29, 2010 at 4:34 PM, Sylvain Beucler <firstname.lastname@example.org> wrote:
> > Hi,
> > When testing the radeon driver we'll get in Squeeze, I noticed that a
> > performance regression: 2D operations are currently around 15x slower
> > in the now-default KMS mode (kernel-based mode-setting).
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=583604#30
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=583653
> > This means that all non-OpenGL games become unplayable or very choppy.
> > While discussing with the package maintainer, I had the unsafe
> > impression that they were more obsessed about enabling KMS (and
> > stopping non-KMS maintenance) rather than perserving performances.
> > I hope this will get fixed for Squeeze :/