Martin wrote:
> Per Jessen wrote:
> [...]
>> I would tend to agree, except I've been monitoring temperatures (with
>> lm_sensor) and CPU-temp barely goes higher than 55-56C.  I'm sure it
>> would with e.g. 4 copies of mprime, but with 2 it stays in that range. 
> 
> That reporting is only as good as what the sensor measures at the
> instant each measurement is taken. It can only be taken as an 'indicator'...
> 
> 
>>> There is another possibility that's just hit discussion in various
>>> forums... As a test, can you disable Core #2 (the 3rd core as numbered
>>> 0 1 2 3) and try your tests again?
>> I know there is a control register for that, but I think I'd have to
>> write some code to get to it :-) 
>> Unless you know of a (linux-based) utility that will do it? 
> 
> You can add "maxcpus=n" onto the kernel boot line to limit the maximum
> number of CPUs that will be used, so maxcpus=2 would be an interesting
> test if that then uses only cores #0 and #1.
> 
> Are there any BIOS settings to tweak?
> 
> Or does anyone know if a /proc or /sys tweak can enable/disable a CPU in
> Linux?

By the infinite flexibility of Linux! :-)

Try:

echo 0 >> /sys/devices/system/cpu/cpu2/online

to disable cpu2

Try:

echo 1 >> /sys/devices/system/cpu/cpu2/online

to enable oncemore.


>>> Also, can you _underclock_ the system and again retest?
>> Yep, just tried running the memory at 667MHz (normally rated for
>> 800MHz), no change. 
>> I couldn't find a BIOS option that would let me underclock the CPU. 
> 
> Do you need to tweak the memory multiplier to change the CPU clock?...


Let us know what you find.

Good luck,
Martin

-- 
----------------
Martin Lomas
[EMAIL PROTECTED]
----------------
_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime

Reply via email to