[EMAIL PROTECTED] wrote:
> As suggested, I applied Arctic Silver to my new Dell's CPU, and
> thankfully the system didn't blow up. However the temperature drop
> was little to none. (I am aware of the need for a break-in period.)
>
> However, today I discovered that the system has a nVidia
> configuration utility that allowed me to raise the system's fan
> speeds significantly (enough to be plainly audible). That caused a
> real temperature drop according to SpeedFan. Under full load on short
> FFTs, the core temps are now typically 49C to 55C, which is at least
> 10 degrees cooler than before. This should be a good thing, right?
That is indeed a good thing. It is, however, still quite a bit hotter
than my machine runs. My machine reports 47.5 degrees C when running
under my desk which has a table cloth hanging down and touching the
case. That's running mprime, not when idle.
> BUT the system still gets fatal errors in the Prime95 torture test,
> though only on the longer FFT lengths. I've had errors at 896K and
> above, but I've run close to 24 hours of short (8K to 64K) FFTs with
> no errors.
Your temperature is high, but not so high I'd be blaming
it, now. Anything under 70 degrees C is unlikely to be
causing the problems you see.
> Any ideas about why this is happening?
You may need to reseat your memory and/or other cards.
I second the suggestion to run memtest. You could remove
the memory sticks and only test one at a time. That way
you won't wonder whether the problems span sticks. If
all sticks pass, then reinstall and rerun one more full
pass of all tests. You might have address conflicts between
cards.
Mike
--
p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);}
Oppose globalization and One World Governments like the UN.
This message made from 100% recycled bits.
You have found the bank of Larn.
I speak only for myself, and I am unanimous in that!
_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime
