[Tfug] suspend to ram question

[JD Rogers]

> When I switched my T40 from APM to ACPI a few years back, my battery
> drain during suspend-to-ram increased pretty significantly.  ACPI is
> still pretty much of a mystery to me and I generally don't suspend for
> more than a couple of days at least, so I never tracked down the
> culprit, but the higher discharge was disappointing.

I remember Ranjan having a problem with ACPI, and there turned out to
be some bug where something (video?) was still being powered. There
was a script to check for this. I think this is it:
http://www.thinkwiki.org/wiki/ACPI_sleep_power_drain_test_script
But if I remember correctly, that bug made suspending longer than a
day impossible, so you likely don't have that. In any case, you should
be able to compare with that script.

I'm pretty sure that if things are working correctly, the power drain
should not depend on acpi or apm, since the only thing powered should
be the ram. One thing I feel like I noticed, but have never measured
is that it seems to drain faster now since I added a dimm.

On my t60, I get down to about 11-12 W when idle and the screen dimmed
all the way and unload all the modules I can (sound, cardbus, usb,
ect). I think I was able to get 10-11 if I used the non-3D video
driver, but that has improved enough now, that I don't feel its worth
it. If I don't go crazy unloading modules, I usually hover around
13-14W, or 16-17 with the screen bright. The real killer is
firefox/java/flash which typically runs the CPU at ~7-8 percent, but
that cascades to heat which then forces the fan on, etc and puts me up
to 20-24.

JDR

--
Jeremy D. Rogers, Ph.D.
Postdoctoral Fellow
Biomedical Engineering
Northwestern University