Q6600 failed/crashed after 10hrs prime...

Associate
Joined
7 Sep 2009
Posts
39
Location
Ireland
I'm gutted right now! I had just overclocked it to 3.7GHz. Prime (small FFTs) ran fine throughout
the night and was still going fine this morning until I came in and found that my system had restarted (blue screen according to the error report).... And this was just after about 10 and a half hours of complete stability, no errors!!!? I have never had an overclock fail on me after this long!

Whats the general outlook towards true stability? Usually people say 8 hours is enough, some say 12. Seeing as it ran stable under intense load for such a long period, whats the likelyhood of me ever experiencing serious issues if I were to settle for this overclock and use it 24/7? Can I call it stable?
 
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
That's what I'm thinking to. It seems to be a dodgy situation to be in either way though, as many people would insist that a stress test such as prime should never fail, no matter the length of time it is run for, whereas an equal amount of people are of the opinion (as we are) that failure after a long period of being stressed would not cause any problems under normal use.

I will try out those other stress tests as well, thanks BeeP.

And does anyone know what a bsod, caused by an overclock, rather than a simple failure in prime, actually means? Is there a difference, or one worse than the other?
 
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
Have you tried any other stress testers. ? Prime is really ott in my opinion. Give Intel burn test a try set to maximum and try 10 passes. If you success your stable in the real desktop level in my book. If prime is your thing then for me you got enough hours under your belt there to say its stable.

McT
I have not yet tested with anything elseJust googled there to find out the temp difference to expect if I run the IBT, and it seems they can reach anywhere between 5-10 degrees higher than with prime! I think I will try it, but I'm nervous to say the least. My temps already peak at 69C in prime, so I guess they may reach the high 70's in the Intel Burn Test. I'll definately be leaving the side of the case off for this.
 
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
Well I ran the Intel Burn Test (max stress level/10 times) and it passed. Would have liked to run it for longer but the temps were pushing me to the brink of a heart attack.

So what do you guys think? Does this really mean 100% stability? By the way I only have 2gigs of ram and someone mentioned that this may have lightened the load a bit and skewed the results. Is there any truth to that?
 
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
I just ran 20 passes in the IBT and it passed again. I hope this is not too good to be true. So does anyone know wether or not using only 2gb of ram actually lessens the stress on the CPU during the Intel Burn Test compared to using 4GB+? It's vital I know.

Btw even though the overclock has seemed stable so far I decided to bring the FSB down one notch (from 412 to 411) which I will test with prime overnight to see if it passes without issue this time. 411x9 gives me exactly 3.7GHz so I'm fine with that. I just figured that because prime failed after such a long period, perhaps only the slightest tweak is required to eliminate any instability.
 
Last edited:
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
I would give it a spin through OCCT, it can show you Vcore stability etc.

TBH passing IBT with 20 is stable enough for most (myself included), I'd call it job done after that.

Just as a matter of interest what Vcore do you have in it for 3.7? I'm at 1.375 for 3.6 & might try for 3.7/3.8 soon. I never go above 60 in prime/IBT/OCCT with my TRUE push pull on.
I'll try out OCCT as well then. May as well. And I'm running 1.5v (realtime) through it for 3.7.

Btw, the system bsod'd during prime again, although this time with the slightly lower FSB it last 25mins longer :rolleyes:

Right so it shouldn't matter wether you have 2gb or 4GB ram in your system when running IBT then? Just as long as you use the maximum memory you have free, am I right?
 
Associate
OP
Joined
7 Sep 2009
Posts
39
Location
Ireland
Correct, need to test all availiable memory, I can be stable using 2-4gb RAM on IBT easily but when i up to max memory (10.5gb availiable in my case) it can fail very quickly
Okay that's good to know because that means my 20 passes of IBT were legitimate! I have this person who swears that you must have more than 2gb of total system memory... That just wouldn't make sense at all. Thanks.

For the two or three runs I have done of the IBT it used about 1100-1200mb of free memory (with maximum stress level). I checked in the task manager and it seemed that it was in fact using as much memory as possible.
 
Back
Top Bottom