On Monday 16 July 2007, T?r?k Edvin wrote:

> On 7/16/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> > First, I noticed that my CPU consumes considerably more
> > power (an extra 0.5W) when running at the lowest speed
> > (1.00GHz) as opposed to when it runs at its native speed
> > (2.16GHz); it may be a hardware flaw (ACPI?) or anything
> > else, I don't know.
>
> What driver do you use to modify the clock?
> If its clock-modulation, its not good for power consumption (only
> for thermal emergency), afaik.

I've tried both acpi and centrino and none of them seems to give any 
power savings at all (see below).

>
> > I wonder
> > whether cpufreq, in as cool and sexy as it is, does make any
> > sense at all (given the fact that it does introduce some
> > overhead). For me, it seems that it doesn't (even when I
> > adjust the minimum frequency so as to avoid the 1.00GHz
> > thing described above). My guess here is that a faster clock
> > allows the CPU to do its work quickly, and then sleep
> > longer. I may be wrong, and I would like to know whether
> > other people experience the same situation.
>
> Try the ondemand governor. It should make the CPU go max speed, do
> the work fast, then go back to low-speed&sleep state.
> The cpuidle patch should also help make better use of C-states.

I did -- theoretically, I agree that a lower speed and a lower voltage 
should produce less heat, and consume less power, just that I 
couldn't measure any reduction in power consumption, so I'm wondering 
whether other people experience the same situation. are you using 
cpufreq? do you get much power savings? I think I will also be 
looking at linux-phc (https://www.dedigentoo.org/trac/linux-phc/) 
when I've got some time.

>
> > An idle kernel, for me,
> > produces (still in runlevel 1) just ~1 wakeup/second (with
> > C3 residency times of ~850-1000ms).
>
> That should be good enough ;)

well, yes, it's good, but not enough: my system still draws about 
17-18 watt on idle (a mostly black tty in runlevel 1, only necessary 
kernel modules, filesystem mounted readonly, no audio, etc.), whereas 
mac os x gets away with as little as 11 (full functionality, all 
drivers, and graphic environment). so I'm wondering what causes such 
a big difference...

> > The next steps will be to identify userspace code that
>
> Try some other hardware tweaks, such as backlight, tv-out-disable,
> etc.

backlight was dimmed down to a very low level during tests. x11 wasn't 
even loaded (though the difference between a linux console and 
x11+vesafb is negligible in my case). everything that wasn't strictly 
necessary was removed from the kernel, and later loaded as a module to 
confirm whether each particular module causes increase in wakeups 
and/or power consumption (an idle snd-hda-intel alone, for example, 
is responsible for a couple of watts in my case).

andrea.
_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power

Reply via email to