Page 1 of 1

Unstable Machine reporting

Posted: Sun Aug 31, 2008 11:12 am
by grumpydaddy
I fold with a gx2 and a 9800gtx both are factory overclocked then further overclocked by myself. whilst one side of the gx2 and the gtx show to be rock solid the other side of the gx2 seems to throw out an unstable machine error once or twice every 24-36 hours.

I realise that the production process will give this variation in quality and that maybe underclocking will resolve this but I am wondering if there are codes reported (for example corestatus=xxx) that point to the weakness when working on seemingly various WU's.

I had thought this was temperature related where the failures were after midday when ambient was higher but I have found some at 5am local which seems to quash that idea.

I really want to know if errors reported contain info that will help me reduce just the offending overclock as you will appreciate this is something of an intermittant fault.

Re: Unstable Machine reporting

Posted: Wed Sep 03, 2008 5:30 am
by bruce
I don't know of any error codes that would be helpful in deciding how stable your machine is.

Temperature is certainly one possible factor, but power supply limitations often turn out to be just as or even more important. What happens to your 12v rail during processing?

Re: Unstable Machine reporting

Posted: Wed Sep 03, 2008 4:28 pm
by grumpydaddy
I fitted a 1kw single rail psu from pcpower & cooling which at 50% load seems to stay rock solid (within .01v) from idle to full load.....The more time passes the more I am convinced this IS temp related especially as with the cooler weather this week the frequency of EUE's has dropped to 0 (I took 18Mhz out of the core too 'cos I' m sure EUE's don't help the science)