is there a bug with prime95?

2500mobile @ 2.3 ghz, 1.725 vcore, 2.8 ram volt, pc3200 ram

[Wed Jul 28 20:20:22 2004]
Self-test 1024K passed!
[Wed Jul 28 20:35:31 2004]
Self-test 8K passed!
[Wed Jul 28 20:50:53 2004]
Self-test 10K passed!
[Wed Jul 28 21:07:52 2004]
Self-test 896K passed!
[Wed Jul 28 21:24:37 2004]
Self-test 768K passed!
[Wed Jul 28 21:43:26 2004]
Self-test 12K passed!
[Wed Jul 28 22:01:56 2004]
Self-test 14K passed!
[Wed Jul 28 22:19:09 2004]
Self-test 640K passed!
[Wed Jul 28 22:34:46 2004]
Self-test 512K passed!
[Wed Jul 28 22:53:35 2004]
Self-test 16K passed!
[Wed Jul 28 23:12:40 2004]
Self-test 20K passed!
[Wed Jul 28 23:29:34 2004]
Self-test 448K passed!
[Wed Jul 28 23:46:52 2004]
Self-test 384K passed!
[Thu Jul 29 00:04:21 2004]
Self-test 24K passed!
[Thu Jul 29 00:22:48 2004]
Self-test 28K passed!
[Thu Jul 29 00:40:41 2004]
Self-test 320K passed!
[Thu Jul 29 00:57:20 2004]
Self-test 256K passed!
[Thu Jul 29 01:15:02 2004]
Self-test 32K passed!
[Thu Jul 29 01:33:05 2004]
Self-test 40K passed!

notice that prime95 never ACTUALLY reported a failure, but the computer rebooted itself.

now @...

2500mobile @ 2.2ghz, 1.725 vcore, 2.8 ram volt, pc3200 ram.

[Thu Jul 29 10:18:00 2004]
Self-test 1024K passed!
[Thu Jul 29 10:33:47 2004]
Self-test 8K passed!
[Thu Jul 29 10:49:51 2004]
Self-test 10K passed!
[Thu Jul 29 11:07:27 2004]
Self-test 896K passed!
[Thu Jul 29 22:55:42 2004]
Self-test 1024K passed!
[Thu Jul 29 23:11:43 2004]
Self-test 8K passed!
[Thu Jul 29 23:27:52 2004]
Self-test 10K passed!
[Thu Jul 29 23:45:31 2004]
Self-test 896K passed!
[Fri Jul 30 00:02:53 2004]
Self-test 768K passed!
[Fri Jul 30 00:18:17 2004]
Self-test 12K passed!
[Fri Jul 30 00:33:21 2004]
Self-test 14K passed!
[Fri Jul 30 00:51:12 2004]
Self-test 640K passed!
[Fri Jul 30 01:07:22 2004]
Self-test 512K passed!
[Fri Jul 30 01:22:27 2004]
Self-test 16K passed!
[Fri Jul 30 01:38:00 2004]
Self-test 20K passed!
[Fri Jul 30 01:55:25 2004]
Self-test 448K passed!
[Fri Jul 30 02:13:20 2004]
Self-test 384K passed!
[Fri Jul 30 02:31:41 2004]
Self-test 24K passed!
[Fri Jul 30 02:46:44 2004]
Self-test 28K passed!
[Fri Jul 30 03:05:13 2004]
Self-test 320K passed!
[Fri Jul 30 03:22:24 2004]
Self-test 256K passed!
[Fri Jul 30 03:41:02 2004]
Self-test 32K passed!
[Fri Jul 30 04:00:04 2004]
Self-test 40K passed!

notice again that prime95 never ACTUALLY found an error, the system simply rebooted. but NOTICE that the system rebooted on the exact same test, and near the same time.

id say prime didnt fail your correct,windows or the machine failed.
most likley is heat on the cpu to high or not enough voltage.
small chance its the ram but if the cpu isnt the issues it may be second in line.

so not enough voltage as in the vcore or is it cause my psu is crappy in general?

oh and i don't think the cpu overheated, cause the load temp is around 50.

Sounds like a PSU problem to me, if it's just a reset.

god damn, i got a 420watter from thermaltake. is this not enough for a simple oc like this??

im still saying the cpu needs more voltage.
the psu should be fine unless your seeing large swings in teh lines in mbm5 during priming.

thenerd said:

god damn, i got a 420watter from thermaltake. is this not enough for a simple oc like this??

Click to expand...

Language language!

What motherboard do you have? What video card? How many PCI cards? How many drives? 420 watts doesn't always go a long way.

Though I am just a newbie to the forums, I would have to disagree with deathstar. The fastest way to make prime find an error is to run at too low of a voltage. I have found that when nearing the limits of your overclock, that in order to not spit out errors, prime usualy needs .1 to .2 volts more than is needed to boot up or game.

Prime does not check for errors every second due to the size of the calculations. I would think that it is possible that your motherboard could suddenly undervolt and cause the computer to reboot between prime's error check. Furtheremore, when I watch prime working, I notice that there are temp spikes corresponding to when prime switches calculations. This would be caused by a power spike and therefore a mini voltage dip. The stress that prime puts on your cpu varries with FFT length, slightly. It might be possible that the transition to whatever comes after 40k FFT's - 192k? - is putting your cpu over the edge and prime does not record the error because it reboots before it gets to the first check.

In this case deathstar would be technically correct but not in the way he was thinking. I wouldn't go higher than 1.725 anyway.

If I were you, I would want to find out what FFT length is causing the reboot. I would bet money that it is either 192k or very close. That seems to cause the highest power consumption on prescotts, northwoods, cellies, p3's, p4 M's and barton mobiles, which leads me to believe that it does not depend on the cpu. Whatever it is, I would then set prime to that FFT length, fire it up after letting the cpu idle for a few minutes, and moniter your voltage and temp until it either crashes or passes. If you find a voltage dip, you're just going to have to live with a computer that is gaming stable but not prime stable, or cut back a little.

If it is a voltage dip, I do not think you have a problem. 100 to 1 odds say that prime is the only thing that will ever cause it to reboot. The only thing you have lost is the prime stable bragging rights.

Final note: Yes my current rig is an intel, but that doesn't mean that I am not familiar with AMD.

Got a possible solution for you. This is a question that goes out to all the people who get these kind of errors. Are you running Bit torrent while using Prime 95?

I have noticed that whenever I have bit torrent running and I try to run Prime95, I get these kind of errors. (Where Prime95 will the pass the main tests, but it won't break down into seperate tests). However, it will work fine as the same OC if I don't have a bit torrent file Downloading.

Btw, if you don't know what bit torrent is, google it, it's good stuff.