[sugar] some new matchbox rpm testing results

Chris Ball cjb
Wed Mar 26 00:18:37 EDT 2008


Hi Martin,

   > It looks like the memory is in Xorg[2], and is around 2 MiB per
   > activity[3]:

First of all, thanks so much for doing this profiling, it's awesome.

I'm unclear on whether this 2 MiB per activity is additive or not --
matchbox seems to destroy the composition pixmap on client window hide,
which suggests it wouldn't be:

   comp_engine_client_hide(Wm *w, Client *client)
   {
   ...
     if (client->picture)
       {
         XRenderFreePicture (w->dpy, client->picture);
         client->picture = None;
       }

Assuming this is additive in the measurements you took, though, is there
any reason that it needs to be?  What composited operations do we want
to be able to perform?  If we're only really trying to speed up things
like overlaying windows or the frame on top of a base window, can we get
away with asking matchbox to only composite on top of the currently
shown window?

   > After launching etoys (which didn't seem to want to draw itself,
   > but dunno if that's my build or the compositing) and stopping it,

I see the following in one of the latest matchbox commit messages, might
be relevant:

"compositor now compiles again - but dont work 100%"

Michael mentioned that etoys doesn't work for him either, and that
Record is failing to render its Xvideo window sometimes too.

Thanks!

- Chris.
-- 
Chris Ball   <cjb at laptop.org>



More information about the Sugar-devel mailing list