Announcement
Collapse
No announcement yet.
The Linux Power Problem Is Widespread, Hits Desktops Too
Collapse
X
-
Originally posted by DebianAroundParis View PostWith 2.6.38 I experience major screen corruption. I regularly have to reboot into 2.6.37.
I wonder if it may be due to my Intel chipset overheating.
Putting the below in /etc/X11/xorg.conf.d/gpu.conf stopped the crashes and corruption.
Code:Section "Device" Identifier "HD4870" Driver "radeon" Option "EnablePageFlip" "FALSE" EndSection
Leave a comment:
-
With 2.6.38 I experience major screen corruption. I regularly have to reboot into 2.6.37.
I wonder if it may be due to my Intel chipset overheating.
Leave a comment:
-
NO kidding! My 3 core Phenom desktop was powering itself off every once in a while (Kubuntu 11.04). I opened it up and boy, was it hot! I took the opportunity to clean up the fans (the had lots of dust and lint), the ventilation grids on the box, etc. Reseated the CPU cooler. I am not really sure how much of this was related to the new kernel. It's running 10C cooler now.
Anyways, the two first lines in powertop seem suspicious (and consistent with what Michael was saying)
Code:42.2% (986.5) [Rescheduling interrupts] <kernel IPI> 19.0% (444.5) [extra timer interrupt]
Leave a comment:
-
I just got back from vacation and opened my bedroom door with my computer sitting idle running 2.6.38.3 for a few days and a rush of hot air came out. My wakeups really aren't that bad though. Under 100/s in a quick check just now with X and a browser running. It's a pretty well overclocked i7 but runs with ondemand which appears to be functioning correctly. Idle temperatures have been higher than normal lately. No real indications as to why though.
Leave a comment:
-
I have a battery backup I can use, or I can use a laptop. I don't have a power meter though.
Leave a comment:
-
Originally posted by skeetre View PostI'm running ubuntu 10.10 and mint 10.10 both with the 2.6.38-8 kernel. I'd be glad to run some tests and/or test out a different kernel.
Did you encounter the regression already in moving from pre-2.6.38 to 2.6.38 in a measurable manner?
I think the issue should affect most people, but due to limited time, mostly looking for people who are confident they are affected. Also knowing the architecture for when anyone reports a problem is useful.
Leave a comment:
-
I'm running ubuntu 10.10 and mint 10.10 both with the 2.6.38-8 kernel. I'd be glad to run some tests and/or test out a different kernel.
Leave a comment:
Leave a comment: