Uff Da! The Linux Power Bug Even More Mysterious
Phoronix: Uff Da! The Linux Power Bug Even More Mysterious
In continuation of Mobile Users Beware: Linux Has Major Power Regression, what's causing this regression in the mainline Linux kernel between Linux 2.6.37 and 2.6.38 (and still outstanding in 2.6.39) is still a mystery. To some surprise, even when disabling power management on one of the ThinkPads, the problem is still present...
I also experienced strange power consumption behavior on my Thinkpad T510 (Core i7-720M). I noticed that while measuring the total power consumption in idle and in "idle+wwan connected". It turned out, that the notebook actually consumed *less* power (about 0.5-1W, don't remember) when WWAN was connected.
I quickly found out, that the reduced power consumption was caused by the 10 wakeups/s caused by my dialup tool. A python script generating the same amount of wakeups led to the same *decreased* power consumption.
Does anyone else see this issue?
My 2.6.39 based ThinkPad X61T (Intel mobo, Intel CPU, Intel IGP) on Fedora 15 is a little warm while sending its Rhythmbox over GbE to run on my desktop's X server (and my desktop's PulseAudio too), but not unusually warm. When I unplug the power, battery estimate says 4 hours remaining. This is pretty typical for a 2 year old ultraportable laptop that's never had its battery replaced (a brand new battery could pull 5 - 6 hours I estimate; I never managed to get more than 6.5 hours even on Windows with the brand new battery when I first got it. Wifi slaughters power usage regardless of OS. Lenovo advertises up to 8 hours, but that's probably with wifi off, minimal screen brightness, with the occasional keystroke in MS Word.)
I'd tentatively say that it's "always been like that" -- except that one time on Fedora 11 when there was a bug in gnome-screensaver that made it constantly use 100% of CPU ... my battery died during a 90 lecture while taking notes from a full charge because I wasn't paying attention to the battery and wasn't aware of the problem at the time.
The battery life on my notebook increased from 2 to 3 hours, since I upgraded from Lucid to Natty, because CPU wakeups went down to 30% than it was before.
Michael, could you post some powertop output from both kernels? Shouldn't take long.
How is power usage being measured?
How is the power consumption being measured? Is independant hardware being used to measure power passing from plug to device (i.e. killawatt and it's more sophisticated brethern)? Or is a power monitoring feature built into the hardware being used? In the latter case this could be a bug in the power monitor software or drivers.
Have you noticed decreased battery life or anything else that would independantly confirm increased power useage?