[ltp] some power saving on thinkpad

[Henrique de Moraes Holschuh]

On Sun, 20 May 2007, Alexey Fisher wrote:
> The lowest state is C4 in PIT
> C0 1,9%
> C1 0%
> C2 0%
> C3 0%
> C4 98,1%
> 
> Henrique de Moraes Holschuh wrote:
> > On Sat, 19 May 2007, Alexey Fisher wrote:
> >>> the difference PIT with HPET on lowest usage for CPU are conservations
> >>> states,
> >>> for PIT it looks like:
> >>> C0 1,9%
> >>> C1 0%
> >>> C2 0%
> >>> C3 98,1%
> >>>
> >>> for HPET:
> >>> C0 0%
> >>> C1 0%
> >>> C2 36,5%
> >>> C3 40,7%
> >>> C4 22,8%

I get with PIT in 2.6.20 readings that are more like what you show in your
HPET tests... weird.

But there is heavy suspicion that dynamic tics and tickless have detrimental
effects (right now - it is a bug to be fixed) on power conservation, as they
are indeed not helping the kernel go from C2 to C3 or C4.

Where is the HPET patch?  I could try to play with it and see if I get some
measurements as well...

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh