Results 1 to 2 of 2
  1. #1
    Joined
    Jun 2003
    Location
    Uppah Nu Yawk
    Posts
    29

    Question Troubleshooting system stability issues

    I have been running Prime95 and MemTest-86 to assess my system stability in conjunction with overclocking my CPU and memory (see system details below). Here's what I'm seeing:

    For Prime95, I run two instances of the stress test simultaneously, one with small FFT's and the other with large ones. I see failures ONLY in the large FFT test, with the time to failure depending upon the FSB setting. At 205MHz, the failures (fatal error for rounding miscalc) occur quickly, typically within 1 to 3 minutes. If I back down to 200MHz, the failures move out to between 1 and 2 hours of runtime. Better, but not good. My memory timings are set at 10-3-3-2 (the 2x512MB Mushkin Black PC3200 modules I'm using are rated 6-2-3-2, overvolted by 0.1V). I keep the CPU internal clock at or below 2400MHz for this testing, where my mobile XP 2500+ proc seems to be completely stable. The fact that I never see failures on the small FFT test seems to corroborate that the instability isn't in the CPU. I also tried running different memory (2x256MB OCZ PC3200 Enhanced Latency) in the system to see if it affected the results, and I still saw the same basic failure rates on the large FFT stress test. This leads me to believe that the issue may be in my Gigabyte 7N400 Pro2 mobo.

    On MemTest-86, stability seems to be much better, but I still see some failures in the 200 to 205MHz FSB range. Somewhere between 12 and 20 hours, I'll see multiple test failures, predominantly on Test 5, with an occasional Test 4 failure (not more than once in a 20+ hr. period).

    I haven't run either test at FSB speeds below 200MHz, though I started the dual Prime95 test at 195MHz before leaving for work this morning.

    So with this lengthy preamble, my basic question is....how do I positively identify what the source of the system instability is? Are there different test conditions or different tests that I should be trying to pin this down? As I mentioned, I suspect the mobo as the culprit, but I really don't want to replace it and then find out I was mistaken. Thanks in advance for your suggestions and insights....

  2. #2
    Joined
    Jun 2003
    Location
    Uppah Nu Yawk
    Posts
    29
    UPDATE: Ran the dual Prime95 tests at 195x12 for 24 hrs. with no errors or warnings. So at least I've defined a workable "floor" to go from here.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •