[OT] CPU temperature control

[OT] CPU temperature control

[Padraig Brady]

Hi,

I'm using a C3 700MHz CPU underclocked to 466MHz (66MHz FSB),
and @ full load I'm getting the CPU to 63°C in a fanless 1U case.
What I would like is to throttle the CPU back X% if the temperature
exceeds say 50% which I can easily read using lm sensors.
So, what's the best way to do this? user space / kernel space??
Note the C3 has a suspend on halt (instruction) option which will
help things also.

cheers,
Padraig.

Re: [OT] CPU temperature control

[Dave Jones]

On Fri, 24 Aug 2001, Padraig Brady wrote:

> I'm using a C3 700MHz CPU underclocked to 466MHz (66MHz FSB),
> and @ full load I'm getting the CPU to 63°C in a fanless 1U case.
> What I would like is to throttle the CPU back X% if the temperature
> exceeds say 50% which I can easily read using lm sensors.
> So, what's the best way to do this? user space / kernel space??
> Note the C3 has a suspend on halt (instruction) option which will
> help things also.

Russell King has been working on a architecture independant framework
for this kind of feature. I recently started work on VIA support for it.
After this weekend, my socket 370 motherboard should turn up and
I'll be able to test / finish the support for it.

You can find the code so far at..
cvs -d :pserver:cvs@pubcvs.arm.linux.org.uk:/mnt/src/cvsroot login
cvs -d :pserver:cvs@pubcvs.arm.linux.org.uk:/mnt/src/cvsroot co cpufreq

So far, there are several implementations in there, in various stages
of completion. The various x86 types need finishing, and it'll be
ready (as long as rmk has nothing else to add to it) for submission.

regards,

Dave.

-- 
| Dave Jones.        http://www.suse.de/~davej
| SuSE Labs

Re: [OT] CPU temperature control

[Russell King]

On Fri, Aug 24, 2001 at 06:06:00PM +0200, Dave Jones wrote:
> So far, there are several implementations in there, in various stages
> of completion. The various x86 types need finishing, and it'll be
> ready (as long as rmk has nothing else to add to it) for submission.

I don't have anything to add to the generic layer; it needs a little
cleanup to remove the "this is debug only" comments against the proc
interface.

--
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html