Thinkpads Forum

hi,
i'm just configuring my new t60 and i'm not quite happy with the power consumption in battery mode.

untill now i'm using cpufeq with the powersave governor and laptop-mode.
i get about 2:30h battery life which is not very much compared to about 4h in windows.

i also tried "aticonfig --set-powerstate=1" (t60 with x1300). this decreases the graphics performance but does not reduce the power consumption in my case (latest fglrx version in debian unstable).

maybe someone knows some tricks how to get ati's powersaving working or has other ideas how to get more battery life?

furthermore it would be nice if other t60 users can post the output of:
cat /proc/acpi/battery/BAT0/state | grep present
when running on battery.

with debian i get about 20000mW.
in windows its about 16000mW (measured with some shareware tool - don't know the name)

thanks

debian unstable
kernel 2.6.17
fglrx 8.24.8

Hi fusch,

I am running Debian / testing (Etch) wtih kernel 2.6.17.4, fglrx 8.26.18.
After setting the GPU into powerstate "1" using aticonfig, I get these values from a cat /proc/acpi/battery/BAT0/state | grep present when running on battery power only:

present: yes
present rate: 19047 mW
present voltage: 12132 mV

That's the rate at about 1/2 brightness on the display. If I drop the brighness to its lowest setting I get about 18945 mW.

Also, cpufreq is being used with powersave governor running in "-l" mode.

I haven't done a rigorous timing of the running life of the machine under various conditions yet, but the battery life monitor in Gnome reports a few minutes shy of 3 hours remaining after a good long charge, for what that's worth.

Is your 4-hour uptime value from Windows an actual measurement you took or is it the value reported by some software utility?

Just thought I'd contribute some info to the effort here. Hopefully we can effectively stretch our machines untethered lifetimes a ways. I have so far been rather unsettled by the behavior of my (first and second, had the planar card replaced already because the temps spooked me) GPU. Perhaps this temp thing is an indicator of poor power saving as well?

I'll keep poking at this as often as I can and continue to help build up a more accurate picture of what these GPUs are actually doing inside our machines. Eventually if we get enough information together we might be able to come to some concrete assertions or, better yet, some solutions or improvements.

--Omou.