Poor X.org performances in Hardy Beta

[Lionel Dricot]

I've installed a fresh Hardy Beta on my Toshiba U200 laptop ( https://wiki.ubuntu.com/LaptopTestingTeam/ToshibaSatelliteU200-163 ) with an intel graphic chipset.

In Gutsy, all was working normally with since Hardy, I experience very poor performances :

- the X.org process always use a few percent of the CPU according to top
- redrawing a window is "visible" (and can take a few seconds when the CPU is busy)
- switching between desktops produces a visible lag
- playing a video with totem-gstreamer produces some slight frames dropping (and consume the whole CPU)

Two bugs seems to be related to this issue :

- scrolling in webpage is very very slow : bug #206055
- Sometimes, when doing something a little bit hard for the graphic card (displaying a complex webpage, pressing a widget, displaying a calendar in evolution), X will simply be killed along with all the TTY process : bug #206049

[Byron (byron-impulzus)]

Confirmed.
I reported the same here : https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/205070

[unggnu (unggnu)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. This particular bug has already been reported and is a duplicate of bug #177492, so it is being marked as such. Please look at the other bug report to see if there is any missing information that you can add, or to see if there is a workaround for the bug.
Feel free to continue to report any other bugs you may find.

[Lionel Dricot (ploum-deactivatedaccount)]

I don't believe that this bug is a duplicate because I don't use Compiz. My report is a fresh Hardy with no desktop effects at all (and thus no-compiz, metacity as the WM)

[unggnu (unggnu)]

Afaik the performance issue happens without compiz too since it is an EXA issue. Please test the Intel driver version 2:2.2.1-1ubuntu6 mentioned in the bug report or wait some time since the fix was uploaded today. If it still happens with this driver remove the duplicate link or post it here.

[Lionel Dricot (ploum-deactivatedaccount)]

Ok, this bug is indeed a duplicate. Thanks :-)