hybrid-graphics-linux team mailing list archive

[Michal Kottman <k0mpjut0r@xxxxxxxxx>]

On Wed, 2011-04-06 at 11:20 +0100, Pete Goodall wrote:
> Hi all,
> 
> So I checked out the latest acpi_call from git, ran make and inserted
> the module.  When I run the test_off.sh script I find a method that
> works.  However, my battery time remaining goes *down* not up.  I also
> have a modified version of the script that basically only tries the
> method that worked for my laptop (\_SB.PCI0.PEG1.GFX0._OFF) and I
> replace OFF with ON.  Afaict this will turn the discrete graphics back
> on.  When I do this my battery life now shows an additional two hours
> remaining.
> 
> This makes no sense to me, so I'm wondering if I'm misunderstanding
> the purpose of the script.  I thought this script was designed to turn
> off the discrete graphics chip so I had less battery drain.  However,
> from this testing it seems I get more battery drain when it is off.
> Am I wrong?

This is VERY strange. It seems that the producer your laptop has a sense
of humor. We generally do not know what the methods do (and don't have
the hardware to try it out, that's why we need your help), and expect
they work according to their name.

In your case, GFX0._OFF, we would expect that it would turn the card off
and the resulting rate would lower. It would be best if you could
provide a listing showing the following commands, to verify your claims:

$ cat /proc/acpi/battery/*/state
$ echo '\_SB.PCI0.PEG1.GFX0._OFF' > /proc/acpi/call

... wait for a while ...

$ cat /proc/acpi/battery/*/state

If it really turns out that _OFF uses more power, then we'll change the
call to _ON :)