 |
|
| Linux Forum Index » Linux - Gentoo Forum » extra timer interrupts since Xorg 1.5?... |
|
Page 1 of 1 |
|
| Author |
Message |
| Vitus Jensen... |
 Posted: Sat Apr 18, 2009 8:29 am |
|
|
|
Guest
|
Hello all!
Since the last world update, which included Xorg 1.5 I'm seeing a lot of
extra timer interrupts in powertop and worse: 30% of the time the CPU is
only in state C2, not C3. This sums up to 3 extra Watts while idle :-(
Does someone notice the same thing? Any idea which program might be
responsible? I've tried the latest unstable x11-drivers/xf86-video-ati
and sys-apps/hal but the situation is still the same. Next step will be
x11-drivers/xf86-input-evdev.
But perhaps this the wrong track and some other package is responsible?
Bye,
Vitus
--
Vitus Jensen, Hannover, Germany, Earth, Universe (current) |
|
|
| Back to top |
|
|
|
| Arthur Hagen... |
| Posted: Sun Apr 19, 2009 5:36 am |
|
|
|
Guest
|
Vitus Jensen <vitus at (no spam) alter-schwede.de> wrote:
Quote: Hello all!
Since the last world update, which included Xorg 1.5 I'm seeing a lot
of extra timer interrupts in powertop and worse: 30% of the time the
CPU is only in state C2, not C3. This sums up to 3 extra Watts while
idle 
Does someone notice the same thing? Any idea which program might be
responsible? I've tried the latest unstable
x11-drivers/xf86-video-ati and sys-apps/hal but the situation is
still the same. Next step will be x11-drivers/xf86-input-evdev.
But perhaps this the wrong track and some other package is
responsible?
You may be on the right track, but weren't there also some recent changes to
eventd (and its hal counterpart)? That might cause these kinds of changes
too.
Regards,
--
*Art |
|
|
| Back to top |
|
|
|
| Vitus Jensen... |
| Posted: Mon Apr 20, 2009 4:59 am |
|
|
|
Guest
|
On Sun, 19 Apr 2009, Arthur Hagen wrote:
Quote: Vitus Jensen <vitus at (no spam) alter-schwede.de> wrote:
Since the last world update, which included Xorg 1.5 I'm seeing a lot
of extra timer interrupts in powertop and worse: 30% of the time the
CPU is only in state C2, not C3. This sums up to 3 extra Watts while
idle
Does someone notice the same thing? Any idea which program might be
responsible? I've tried the latest unstable
x11-drivers/xf86-video-ati and sys-apps/hal but the situation is
still the same. Next step will be x11-drivers/xf86-input-evdev.
But perhaps this the wrong track and some other package is
responsible?
You may be on the right track, but weren't there also some recent changes to
eventd (and its hal counterpart)? That might cause these kinds of changes
too.
Yes, there were hal updates and the kbd/mouse interface changed etc. A
very complex change when it comes to find new wakeups. But the effect I'm
seeing may even be related to wifi, at least it vanished after unloading
ath5k. And didn't come back after readding :-(
Today I'm at
C0 (Prozessor läuft) ( 1,3%) 1,74 GHz 0,0%
C1 0,0ms ( 0,0%) 1333 MHz 0,0%
C2 19,7ms (26,5%) 1066 MHz 0,0%
C3 21,1ms (72,1%) 800 MHz 100,0%
With 26% extra interrupts. I will first try to get 2.6.29 running on my
thinkpad before looking at it again. Which is a task by itself because it
needs a patch to get apci=rsdp working. Oh well...
Vitus
--
Vitus Jensen, Hannover, Germany, Earth, Universe (current) |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Fri Mar 12, 2010 3:46 pm
|
|