[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: [PATCH] powerdevil-remove-cpufreq.patch
From:       Sebastian =?iso-8859-1?q?K=FCgler?= <sebas () kde ! org>
Date:       2010-03-08 14:09:00
Message-ID: 201003081509.01187.sebas () kde ! org
[Download RAW message or body]

On Monday 08 March 2010 14:39:31 Holger Macht wrote:
> >> What numbers are we talking about? Is the machine completely idle when
> >> measuring? If not, what are the workloads, are they comparable?
> >
> > uhm I can't prove it at the moment, since I'm at work, it's empirical 
> > observation. The machine is not idle, I just unplug, set the 'extreme 
> > powersave' profile (or leave the default powersave profile when I forget
> > to  change it), raise the brightness up 3 notches (of ~12), then keep
> > using kate and firefox (typical web development) until the battery
> > almost dries out.
> >
> > I simply noticed i could get some more battery life with "extreme
> > powersave",  rather than simply "powersave", at the same screen
> > brightness. Quantifying exactly that difference is out of my
> > possibilities at the moment, sorry.
> 
> Ok, so you're not only changing the CPU frequency scaling policy but
> rather a whole profile which might do other things, too. Didn't check
> which, but this of course can have some impact on battery runtime. It's
> interesting if there's prove if just and only changing the CPU freq
> governor from ondemand to powersave gives you a lower power consumption.

BTW, a suitable tool to measure power consumption is powertop.

We shouldn't hold back a patch that is the result of the combined knowledge of the 
kernel power management community (which is what Holger's reference to MJG's article 
really is).

The other comments in this thread are unfortunately a lot less informed.
-- 
sebas

http://www.kde.org | http://vizZzion.org | GPG Key ID: 9119 0EF9
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]