Tom Brinkman grabbed a keyboard and wrote:
>
> On Tuesday 19 March 2002 13:13, David Guntner wrote:
>>
>> Those sound useful. :-) Are they on one of the CDs, or do they
>> have to be downloaded from somewhere? Currently, the only way I can
>> check my CPU temp is to reboot and then go into the BIOS - not the
>> most optimal way... :-)
>
> The temp displayed in bios is for the most part useless. The
> system hasn't heated up yet, and HLT's (halts) aren't being sent by
> the kernel to the cpu. So while there's little load, there's no
> idling either. Cpu core temp can rise and fall dramatically with load.
> Shut down for some time, I'll see cpu temps from bios that are just
> above room temp (~30C), but by the time I get to KDE they're in the
> normal mid 40's.
Unfortunately, krellm doesn't seem to display my CPU temprature, so I'm not
sure how I'm going to monitor it. BTW, just so you're aware of it, my
Linux machine (ML 8.1) is *never* powered off unless I'm having to get
inside the case to connect/disconnect something. So the temprature should
be staying pretty stable inside.
I believe you on the CPU temp changing as the load goes up and down, I just
don't know how to actually monitor it while the system is running, since
those utilities don't seem to know where the CPU temp is being reported.
>> I went to AMD's site, and while I could find a PDF document titled
>> "AMD-K6- 2 Processor Data Sheet," the Thermal Design section only says
>> that it's supposed to run with a case temprature within a
>> certain range. They also don't mention the 550MHz chip.
>
> Well, maybe you can get the info somewhere, but it should be on
> their site. FWIW, my 1.4/266 Tbird is spec'd at 95C, but the 1.4 is
> also the hottest cpu AMD makes. Your K6 should be lower, 80C, maybe
> 85C is probly a decent guess on the high end, but it could be in the
> 70's.
At least when I reboot to look at it in the BIOS, it's well below that.
I'll keep looking, but if it isn't in the "Themal Design" section of the
Processor Data Sheet, I'm kind-of at a loss for where to look... :-)
>> I'm kind-of one step ahead of you there. :-) I pulled my 550MHz chip,
>> thinking that the problem could be with the CPU, and replaced it with
>> the old 350MHz chip that the 550MHz chip had replaced
>> originally (I tend not to throw away components when they get
>> replaced :). So the heatsink has a nice fresh coat of heatsink
>> grease on it, and I did blow everything out and reseated several
>> connectors (can't remember if I got them all or not, though...).
>
> Without gettin too far into it, cpu's that are run too hot, too
> long experience 'cpu creep'. There's actual migration and increased
> resistance of the internal components. Once that's happened, there's
> no going back. So it's good policy to monitor temp, and test
> occaisionly with memtest86 and cpuburn. Another benefit of
> monitoring cpu temp is you get early warning of runaway procceses,
> even if the system seems fairly normal in use.
I agree. Again (just for information), the 350MHz chip that's in there now
never gave me any problems. I upgraded to the 550MHz chip to get more
speed, nothing more. I fell back to the 350MHz chip in an effort to see if
the problem was with the CPU. I'm running a different CPU but still having
the same hang problem, which is why I'm kind-of doubting a problem with the
CPU at this time. I'm persuing the testing information that you've been
providing because A) I'd like to be as sure that it's not the CPU as
possible, and B) I find the discussion to be quite interesting and
informative.
Thanks again for all the information and suggestions.
--Dave
--
David Guntner GEnie: Just say NO!
http://www.akaMail.com/pgpkey/davidg or key server
for PGP Public key
Want to buy your Pack or Services from MandrakeSoft?
Go to http://www.mandrakestore.com
