On Saturday 22 September 2007 13:34, Toni wrote:
> Eberhard Roloff wrote:
> > just a very wild guess:
> > I had similar problems years ago on an old compaq laptop.
> > There it helped to use the kernel option "nomce" at the grub command
> > line (at that time, it was lilo, but you get the idea)
> >
> > Just let me know how it works...
> >
> > regards
> > Eberhard
>
> After some 'googling' I realize that this command just disables
> self-diagnosis checks performed on the CPU. Well, in my case the
> messages thouse checks left in the log file do not bother me; this
> seem to be a side (simultaneous) effect, not the cause of the
> freeze of the system.
>
> Rajko:
>
> Thank you for your exhaustive information. I got a copy of that pdf.
> I'm not sure to understand yours remarks (nor my technical nor my
> English levels are gods enough) but as I understand, at 56ºC (at this
> precis moment there is 39ºC) I'm more than safe from the 90ºC "die"
> temperature for the cpu. On the other hand, 0,03V above the ideal
> voltage of the core either seems too exaggerate. In any case, I don't
> know if my old power-source is capable of tune the different output
> voltages.

Toni,

program sensors, that you use to see temperature, can be very precise and way 
out of proper value. 

It depends on settings in 
  /etc/sensors.conf 
and sometimes they are some default taken from sensor chip manufacturer 
recommended or example schematics, not always from mainboard manufacturer. If 
mainboard fabricant used different schematic, values that program sensors 
gives can be completely wrong. You have to check that they are OK. The check 
can be BIOS, if it has such menu where you can see temperature. 

The temperature setting depends on electronic element used to measure 
temperature and it's position. If wrong element is used temperature on 
display can be completely wrong. 

You can see AMD recommendations in 23794.pdf. From that is obvious that duron 
has no internal termperature sensor, but one has to be supplied by 
motherboard manufacturer. They recommend that temperature doesn't exceed 40 C 
if thermal element is positioned as described in pdf. 

That tells, that you have to look in BIOS and check temperature there, see 
motherboard manual if it has any information about temperatures and what they 
should be. 

It can be some software error, but hardware is in this case faster to check 
and eliminate from checklist. Starting with thermal conditions, than hard 
disk. RAM is checked to some extent, as you said in previous mail.

-- 
Regards,
Rajko.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to