Arch Linux

enyaw_ecurb

Member

Registered: 2010-02-12

Posts: 28

[Solved] cpufreq interfering with window moving

Hi there,

I'm posting my issue in Desktop Environments fully aware that it is probably not an issue of my wm but I just need some advice how I can figure out the problem.

I'm running fluxbox on a amd athlon II x4 with a ati hd5770 running catalyst.
If I activate cpufreq using ondemand or conservative governor several window operations slow down considerably. And they not just slow down (this was somewhat expected with crappy catalyst accelerating) but for example window moving just "hangs". If i try to move a window (opaque moving off in fluxbox) in front of a complex surface (like browser window) it really stops moving, following my mouse pointer with fully noticeable lag, leaving a trail of afterimages.
By chance I noticed that this does not happen if I run my cpus on full frequency. Moving windows still drives one cpu to 50% but it does not hang.
Apparently cpufreq interferes with "demanding" (ha, ha) X operations. My question is if this behaviour can be considered a bug of some component or if there is any way to configure my system to behave more responsive. (remark: Easy solution is to deactivate cpufreq but I'm searching for a *better* solution -> yeah I know I'm beyond help )

Any help appreciated!

greets

enyaw_ecurb

Last edited by enyaw_ecurb (2010-05-01 09:54:50)

Mr.Elendig

#archlinux@freenode channel op

From: The intertubes

Registered: 2004-11-07

Posts: 4,092

Re: [Solved] cpufreq interfering with window moving

There is a bug with catalyst and window moving/resizing/fullscreen. You might be experiencing it. See the ati bar and grill thread for details (and a patched xorg that is suposed to work better)

---

Evil #archlinux@libera.chat channel op and general support dude.
. files on github, Screenshots, Random pics and the rest

enyaw_ecurb

Member

Registered: 2010-02-12

Posts: 28

Re: [Solved] cpufreq interfering with window moving

Thanks! That did the trick!

It still tears like hell and drives my cpu to madness but at least the "hanging" is gone. However it seems that sufficient cpu-power can hide the terrible catalyst performance and create the illusion of a responsive desktop. Maybe this is why some people do not experience this "bug"...

I still do not like the idea that I have a graphiccard that is probably faster than my last three pcs taken together but does not provide decent (or any at all) 2d-acceleration. Maybe all will be well when the open-source driver incorporate support for my card. Blast you, ati!

Enough ranting, thank you again!