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

List:       kde-devel
Subject:    Re: [PATCH] powerdevil-remove-cpufreq.patch
From:       Martin Sandsmark <sandsmark () samfundet ! no>
Date:       2010-03-08 19:46:54
Message-ID: 201003082046.54995.sandsmark () samfundet ! no
[Download RAW message or body]

On Monday 8. March 2010 20.05.16 Sebastian Kügler wrote:
> By not offering options that 98% of the users won't understand correctly,
> we're improving the user experience. (The irony is that pretty much
> everybody on this list belongs to these 98% and at the same time claiming
> that the options is vital. :/) I'd suggest to re-read those references
> Holger posted. It's not like they're written by random people, and it's
> also not the case that re-iterating the reasoning in there is efficient,
> or fun.

I've read mjg's blog before, and he often ignores important points.
Like fans ...

Or when he riled on PowerDevil¹ for providing a mechanism to disable 
composited desktop effects, because compositing moves all the processing to 
the GPU, and should be much more effective, according to him. This completely 
ignoring all the software fallbacks that are used in pretty much all graphics 
drivers, and the extra CPU needed just for computing transitions and whatnot.

So if he gets stuff so wrong on this stuff, I am not sure I trust him in other 
areas.


> I'd suggest that those, who those who disagree with Holger's patch, and
> thus claim that pretty much everybody who's working on power management at
> kernel and HAL levels gear up and start collecting data, because unless we
> can *prove* that those options make a difference, we should trust our
> lower level peers and not offer these knobs that hurt more than they do
> good.

I thought it was pretty intuitive that an active component like a fan would 
make a significant difference in battery time.


> Really, kernel power management is advanced enough these days that we don't
> have to switch CPU frequency manually. And that's a good thing, because it
> *just works*.

... if you don't care about your battery time, apparently.


¹: http://mjg59.livejournal.com/101706.html?thread=1169994
-- 
Martin T. Sandsmark
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

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