[Dev] Info regarding X11 is too slow on Lemote Yeeloong
Daniel Clark at GNU
dclark at gnu.org
Mon May 21 00:23:00 GMT 2012
FYI...
---------- Forwarded message ----------
From: Matt Turner <INVALID.NOREPLY at gnu.org>
Date: Sun, May 20, 2012 at 7:49 PM
Subject: [bug #28169] X11 is too slow
To: Danny Clark <dclark at gnu.org>, leio at gentoo.org, "Richard M. Stallman" <
rms at gnu.org>, Robert Millan <rmh.aybabtu at gmail.com>, Sylvain Beucler <
beuc at gnu.org>, Graziano <graziano at inventati.org>, Sam Geeraerts <
samgee at elmundolibre.be>, Matt Turner <mattst88 at gmail.com>, bernie at gnu.org,
brett at fsf.org, Peter Olson <peabo at fsf.org>, Roman Mamedov <
romanrm at member.fsf.org>, zhangm at lemote.com, Karl Goetz <karl at kgoetz.id.au>
Follow-up Comment #12, bug #28169 (project gnewsense):
I've optimized pixman quite a bit with Loongson's multimedia instructions,
using the existing MMX code. See
http://mattst88.com/blog/2012/05/17/Optimizing_pixman_for_Loongson:_Process_and_Results/
They're available in pixman-0.25.6 and will be in the pixman-0.26.0 stable
release.
Further optimization work includes
- Bilinear/nearest scaling compositing functions for pixman, which are
*really* important for Firefox performance
- Optimized xf86XVCopyPacked for Xv performance. Really, I think pixman
should grow support for all these YUV-type formats, and the code in the X
server should go away
- There may be some way to improve XAA performance that's regressed a lot
in
the last few years, but I don't know if this is doable
Then, there's the work of simply getting the upstream X server working on
the
Yeeloong..
It would help me if someone could tell me what the differences are between
the
upstream 1.7.6 siliconmotion driver and the 2.2.8 version -- and why the
changes in the 2.2.8 driver aren't upstream?
_______________________________________________________
Reply to this item at:
<http://savannah.nongnu.org/bugs/?28169>
_______________________________________________
Message sent via/by Savannah
http://savannah.nongnu.org/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.parabola.nu/pipermail/dev/attachments/20120521/74129752/attachment.htm>
More information about the Dev
mailing list