Mailinglist Archive: radeonhd (364 mails)

[JP Fournier <jape42@xxxxxxxxx>]

OK.  Moved to master for radeonhd and still using the r6xx-r7xx-support branch 
for DRM.  Still get high CPU "sometimes".  This is a bit weird.  Sometimes I 
play a 1080i60 recording I get high CPU usage for a while (say 2 minutes):

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 6013 root      20   0  549m 237m 108m R   98  6.3   1:50.18 Xorg
 6875 jape      20   0  380m 237m 119m S   34  6.3   1:23.25 mythfrontend.re

But then it stabilizes:
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 6013 root      20   0  549m 237m 108m R   38  6.3   6:23.96 Xorg
 6875 jape      20   0  386m 243m 119m S   35  6.4   4:30.08 mythfrontend.re

And it isn't 100% reproducible.  Play the same (or another) recording again 
and CPU usage is reasonable, until it's not.

I've seen HDMI audio and video dropouts even in the low CPU usage scenario.

Any ideas?

(I haven't yet tried out Chris' patch)

jp

On Saturday 07 March 2009 15:08:05 Yang Zhao wrote:
> 2009/3/7 JP Fournier <jape42@xxxxxxxxx>:
> > I've continued doing some more testing on the r6xx_r7xx_branch.  Trying
> > to play over the air 1080i60 mpeg2 video isn't working well for me in
> > current snapshots, but if I go back about a month, things do work - and
> > CPU usage is reasonable.  No problems playing lower res videos(640x480)
> > though.  Output resolution is always 1920x1080 over HDMI.
> > ...
> > root@vcr3:/home/jape/xf86-video-radeonhd# git log
> > commit 6f201b626aa31548537a968c13f073a7a02911f3
>
> r6xx-r7xx-support was merged into master recently, so you should be
> testing that instead. That commit, in particular, was not merged into
> master.
>
> Cheers,

--
To unsubscribe, e-mail: radeonhd+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: radeonhd+help@xxxxxxxxxxxx