Created attachment 19621 [details] XOrg log with intel driver 2.4.2 After updating xf86-video-i810 1.7.4 to 2.4.2 various redrawing operations are slower. And some application (only xchat here crash with an xlib error). I'm using X.Org X Server 1.4.2 with XFCE and Composite enabled. The XChat error goes away when "NoAccel" is being used. Of course with "NoAccel" everything gets unusable slow (even drawing simple resize rectangles). EXANoComposite is a bit better at those resize rectangles but drag and droping of windows is extreemly slow. I'm using transparency while moving windows so unaccelerated composite really hits performance here. Even without disabling acceleration it's slower than with 1.7.4 (Window operations seem to use more CPU) GFX Card: 00:02.1 Display controller: Intel Corporation 82Q963/Q965 Integrated Graphics Controller (rev 02) System: Linux grumpy 2.6.23-gentoo-r9 #1 SMP Wed Apr 16 10:47:51 CEST 2008 x86_64 Intel(R) Core(TM)2 CPU 6400 @ 2.13GHz GenuineIntel GNU/Linux The XChat error with accelerated composite is a generic xlib one: The program 'xchat' received an X Window System error. This probably reflects a bug in the program. The error was 'BadImplementation (server does not implement operation)'. (Details: serial 313 error_code 17 request_code 146 minor_code 5) (Note to programmers: normally, X errors are reported asynchronously; that is, you will receive the error a while after causing it. To debug your program, run it with the --sync command line option to change this behavior. You can then get a meaningful backtrace from your debugger if you break on the gdk_x_error() function.) "MIT-SHM (opcode: 146, base event: 78, base error: 166)" tells me 2.4.2 could probably work here if i disable that extension, but i guess that also lowers performance with some pixmap handing. I'm attaching an Xorg log for each (1.7.4 and 2.4.2) startup
Created attachment 19622 [details] XOrg log with intel driver 1.7.4
Created attachment 19623 [details] xorg.conf for 2.4.2
> The program 'xchat' received an X Window System error. > This probably reflects a bug in the program. Indeed, it does; xchat doesn't check whether SHM pixmaps are supported (they aren't with EXA for performance reasons). I submitted a fix for this in xchat-gnome a while ago. You should be able to avoid this by not using a fake translucent background in xchat.
Could you give me that patch? I'm maybe able to adapt it for "vanilla" X-Chat.
(In reply to comment #4) > Could you give me that patch? I'm maybe able to adapt it for "vanilla" X-Chat. See http://bugzilla.gnome.org/show_bug.cgi?id=398315 .
So the xchat crash is explained above. What's left is that some things are slower. We've addressed some performance issues in the 2.5.0 driver release. I'd be interested to know if you could try it and report back if you notice any difference: http://intellinuxgraphics.org/download.html Thanks, -Carl
ping test result update...
Sorry, result will follow during the next 2 Days. I was out of the Office the past 2 Weeks and had no access to this box.
I'm running 2.5.0 now and it's still alot slower than i810 1.7.4 for many operations. The most noticable is: Slower virtual desktop switching. It can take 0.5s - 1s until everything is redrawn and the desktop usable again. Tab switching in Firefox. It's almost like with the virtual desktops. It takes alot longer until the tab is updated and usable. Opening a new Firefox window (while it's running already) also seems slower. Hard to describe but it overall feels a bit like the machine is under heavy load but it's not. The CPU is totaly idle when i start trying out things. Composite alpha rendering is really fast though. A huge "Terminal" with a true transparent background (50% shaded, composite) is fast when moving the window around on top of others. This seems faster than with 2.4.x.
Are you still having problems with the current 2D driver using UXA on a 2.6.29 or newer kernel? It should fix many 2D performance issues, with virtual desktop switching being a big one.
Sorry, but i no longer have access to this piece of hardware (last job). I'm unable to verify this anymore. Feel free to close this bug.
(In reply to comment #11) > Sorry, but i no longer have access to this piece of hardware (last job). I'm > unable to verify this anymore. Feel free to close this bug. Thanks for your report. I'm sorry we weren't more timely, but hopefully things are actually fixed now. -Carl
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.