radeon EXA performance questions
Simon Thum
simon.thum at gmx.de
Wed Jul 30 14:01:40 PDT 2008
Tiago Vignatti wrote:
> Simon Thum escreveu:
>> I agree in general, but with EXA I seldom experience complete
>> drop-outs where not even the mouse cursor moves. Seldom (1 per 5-15
>> min or so, depending on whatever), but not present with XAA.
>>
>> I use to recent git server/driver, EXA, R200.
>
> My first guess is that your cursor is being sw rendered. Can you confirm
> this on log please?
(II) RADEON(0): Render acceleration enabled for R200 type cards.
(II) RADEON(0): Setting EXA maxPitchBytes
(II) EXA(0): Offscreen pixmap area of 128040960 bytes
(II) EXA(0): Driver registered support for the following operations:
(II) Solid
(II) Copy
(II) Composite (RENDER acceleration)
(II) UploadToScreen
(II) DownloadFromScreen
(II) RADEON(0): Acceleration enabled
(**) Option "dpms"
(**) RADEON(0): DPMS enabled
(==) RADEON(0): Silken mouse enabled
Seems not to be the case. But actually I didn't pay much attention on
getting the setup right, it might have had to do with my spontaneus
Composite extension experiments - I could not reproduce it anymore...
sorry for the noise!
@Michael: It was usually ~300 ms. Could font rasterizing cause such a
delay? It occurred when starting apps or showing windows.
EXA improved a lot, but still, XAA is faster especially for focus
changes and displaying windows (e.g. maximizing or switching to a
complex app takes 1/3 sec or even more). Obviously some ops are still
suboptimal. Is there interest in oprofile data?
I have attached a patch I like to include personally, it fixes the XAA
issues for my system. With the patch, the firefox issues are gone and
consoles are useable too. It would of course only make sense with
Offscreen pixmaps and XAA enabled. Maybe it helps the lost souls stuck
with xserver < 1.5 for some time.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: xf86-video-ati.patch
URL: <http://lists.x.org/archives/xorg/attachments/20080730/daf6a6b2/attachment.ksh>
More information about the xorg
mailing list