1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

What test(s) to run to ensure stability of OC?

Discussion in 'AMD Overclocking' started by Mark Smith, Nov 8, 2003.

  1. Mark Smith

    Mark Smith Guest

    The subject says it all:
    What test(s) to run to ensure stability of OC?

    Have my 2500 Barton on my ASUS A7N8X running at 10.5x with 200 multiplier,
    using pc2700DDR memory, seems fine at 2100 with stock cooling and one case
    fan (CPU at 39 degrees, MB at 18 ambient about 65 (I like it cool!)).

    Want to make sure it truly is stable and error-free at this rate, then maybe
    bump it up again????

    ps,

    if test is tricky (i.e. Memtest86 must be run from floppy), please give me a
    tip or two,

    Thanx,

    Mark
     
    Mark Smith, Nov 8, 2003
    #1
    1. Advertisements

  2. Mark Smith

    dr_hardware Guest

    First, run http://www.memtest86.com/ to test your memory. Then run Prime95
    http://www.mersenne.org/freesoft.htm for at least 12 hours. If it runs
    Prime95 without errors for that length of time, it's stable as a rock.
     
    dr_hardware, Nov 8, 2003
    #2
    1. Advertisements

  3. My general method is keep bumping up the fsb until memtest86 (you need a
    floppy for this) fails early on (within the first 5 mins or so). Then, I
    back off the FSB until memtest86 will complete a full run without any
    errors. Then, boot into windows (should have no problem doing this) and fire
    up prime95 into stability test mode (the middle setting, not the
    consume-all-ram setting, as your memory should be ok). Let it do this for an
    hour or so (or until it locks up/reports an error), and if it locks up (or
    reports an error) either back off the multiplier or fsb. If it goes fine,
    increase the multiplier until it breaks (and optionally back off on the fsb
    a little if you want CPU grunt instead of bandwidth). Takes a bit of work to
    find the optimal fsb and multiplier though.

    Interestingly, if memtest86 fails during cached tests and not during
    uncached ones, additional vcore may not help! If I take my XP1700 too high
    (too high being above about 1660 MHz, pretty shocking for a AGOIA-9) I get
    errors in the cached test, and upping the vcore just makes it worse. I'm
    guessing there's leakage between the cells ore something, and a higher vcore
    makes it worse. Temperature is not a problem, as the diode stayes under 65
    deg C using CPUBurn.
     
    Michael Brown, Nov 8, 2003
    #3
  4. Mark Smith

    QBall Guest

    Prime 95 is FSB not gross clock dependent.


     
    QBall, Nov 8, 2003
    #4
  5. Not in my experience. Prime95 fails (instant lockup) as 152*11 but runs fine
    (several hours) at 152*10.5 ...

    [...]
     
    Michael Brown, Nov 8, 2003
    #5
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.