I have found that Prime95 is a very sensitive test of the stability of an
overclocked system. Except for two old p100/120 systems I overclock all
three of my other Prime95 machines. I use the following procedure:
1. Set up the machine for whatever clock rate I am going to try. This
involves adjusting Front Side Bus and or core voltage.
2. Boot up into Windows. If this fails go to step 1.
3. In local.ini set SelfTestxxxPassed=0 for whatever size FFT the current
test in worktodo.ini needs.
4. Start Prime95. If this test fails go to step 1 and try a lower rate or
higher core voltage. This test takes one hour.
5. Let Prime95 run with new settings. Check results.txt for indications of
possible hardware failure. If any of these show up go to step 1 and try a
lower rate or higher core voltage. In my experience these errors will show
within a day or two.
Using this procedure I have pushed a P III 600E to 843 MHz. I had it
running at 852 MHz but encountered a few errors after a few days and backed
down to 843 where I have had no errors. The procedure has resulted from
this experience and also with overclocking a Celeron 300A to 464 MHz and a
P II 333 to 400. I plan on using this procedure when I get the cooling
upgraded on the 600E and also when I get another system built up using a
600 or 650 with the new chip stepping.
Remeber just because it boots into Windows and appears to run Prime95 just
fine that does not mean that everything is going right. From what I
understand LL tests that encounter some number of errors are scheduled for
a rerun and as result make no real contribution to the overall effort.
Comments and other approaches are appreciated.
Terry
Terry S. Arnold 2975 B Street San Diego, CA 92102 USA
[EMAIL PROTECTED] (619) 235-8181 (voice) (619) 235-0016 (fax)
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
