[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: metacity slowness



On Mon, Aug 25, 2003 at 01:14:54AM +0200, Michel Dänzer scribbled:
> On Mon, 2003-08-25 at 00:59, Marek Habersack wrote:
> > On Mon, Aug 25, 2003 at 12:31:32AM +0200, Michel Dänzer scribbled:
> > 
> > > > The x11perf is just for comparison with some model system, I presume. 
> > > 
> > > If the bandwidth to the framebuffer is low, that may explain at least
> > > some of the slowness you're seeing (apps using the RENDER extension slow
> > > to handle expose events).
> > A comment in metacity code states that it uses the pixmap backstore to avoid
> > handling the expose events. 
> 
> The same unused code for the minimize effect? :) Or for the decorations
It does seem to be used from effects.c, but not for workspace switches,
indeed.

> maybe? Again, the window manager doesn't draw the app window contents
> for all I know...
Actually, you know, the slowness might be due to the code that draws the
workspace background on switches (in draw-workspace.c). But then, why would
it be slower on debian than on gentoo/slackware? Because of the fact that
debian ships binaries optimized for i386 while the other two (or at least
gentoo) optimize for the higher models?

> > Is it possible that the pixmap -> surface blit might be causing the 
> > slowness because of a 2D acceleration bug in the radeon driver?
> 
> Unlikely. BTW, using this technique (pixmap as window background) is
> recommended by Mr. XAA Mark Vojkovich himself as an optimization to
> avoid expose events.
Hm. So maybe it's the right thing to do...
 
> > > I can't seem to tell much difference between metacity and sawfish 
> > > here (except that opaque window resizing works much better with 
> > > metacity :), YMMV.
> > You mean the stroboscope effect? :)
> 
> What's that? :) I mean it's snappy with metacity.
For me it flickers nicely - the faster I resize, the more flicker on the
window edges - just like with a stroboscope :)

marek

Attachment: pgpRpaWVKu6hI.pgp
Description: PGP signature


Reply to: