2.8Ghz overclock NOT stable?

[Celegorm]

You might have just hit a hole in that' chip's/board's capibilities. While 2.8 may not work, 2.85 might.

 

[Bobert93]

I'm aiming for an extremely stable 3-3.2Ghz clock that will not exceed temps of high 50*s C.
.

you dont have to worry about temps with zalman , one again , mines around 25idle and 40 underload ,thats at 3.4ghz

 

[Kage]

Haha, mines around 45 degrees idle on all 4 cores, but it does get pretty hot in this room

 

[A True Folder]

....... Do you guys think my voltage was still too low?

Yes, a total restart is symptomatic of not enough voltage, next step up after that is threads failing, once you get passed that you have stability

Have a look via whatever software you are using to see what volts the Auto setting is supplying, you then have the starting point if you wish to lower vcore and temps manually.

 

[GhostGT]

Okay, update. Set Vcore to 1.22somethingish (lol) and i'm currently clocked at 2.88ghz. Ran 4 Prime95 tests and two of the instances came up with a FATAL ERROR and stopped. The other two tests kept running but I stopped them.

Temps barely touched 60*.


What does this mean?

 

[A True Folder]

when you say ran four instances of prime do you mean four worker threads?

Try 1.22somethingish + one more incrementtypethingish



Have you altered any other voltages?? NB? VDIMM?

 

[GhostGT]

Update: I raised VCore to 1.2375. That's two more increments to the vcore.

I haven't touched any other settings except vcore and fsb.

Ran 4 instances of prime95 again (as in, copy/pasted folder and ran FOUR individual windows of it, with EACH one running a test).

Again, got myself an error in ONE window.

ERROR: Rounding was .5, expected less than .4.

Do you guys think this has to do with the overclock? Or just because four instances of P95 might be conflicting with each other...

 

[Bobert93]

why dont you just go back to 3ghz like it used to be,
wasnt that overclock stable

 

[GhostGT]

Honestly, my system has been fine up till this point, like you said. BUT I've never run more than one instance of P95. I've used CPU Stability Test before and at 100% load at 3.0Ghz i've had no errors or problems whatsoever. Looks like its only P95 giving me crap about it.

I might just stay at 2.8ghz until I go ahead and order the Zalman.

 

[A True Folder]

If you're running four instances of prime95 (16 worker threads) then I'm not surprised your pc isn't stable, that's a hell of a workload, why would you run four instances?