Re: powerd adaptive mode latching

From: Kevin Oberman <oberman_at_es.net>
Date: Fri, 11 Jan 2008 10:13:44 -0800
> Date: Fri, 11 Jan 2008 11:51:07 +0000
> From: "Igor Mozolevsky" <igor_at_hybrid-lab.co.uk>
> Sender: owner-freebsd-current_at_freebsd.org
> 
> On 11/01/2008, Dag-Erling Smørgrav <des_at_des.no> wrote:
> 
> > The whole point of powerd is to automatically adapt to the load.  If
> > your laptop "lags too much" powerd should detect that and increase the
> > CPU frequency.  However, poor interactive response can result when CPU
> > frequency fluctuates too quickly; perhaps your problem will go away with
> > Andrew's patch, which is intended to dampen such fluctuations?
> 
> I was also thinking that the power increase probably shouldn't be
> linear, maybe there should be a choice of algs: 2-step, 50%-100%, etc.
> Normally you don't mind power dropping slowly, but for most things you
> don't want to wait N-seconds for the clock to be stepped up to its
> full throttle...

The part of powerd.c that controls this is very easy to understand and
modify. I write almost no C these days and I had no problems changing
the adaptive algorithm to try other things.

Try your idea and see how it works.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman_at_es.net			Phone: +1 510 486-8634
Key fingerprint:059B 2DDF 031C 9BA3 14A4  EADA 927D EBB3 987B 3751

Received on Fri Jan 11 2008 - 17:13:49 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:25 UTC