Testing OC with Orthos

[GSte]

I've been trying to test my CPU overclock with Orthos. 2 quick questions:

Which test should I use, just the CPU small FFTs?

How long should I run it for?

Cheers

[Clunk]

Use the blend test.

I leave it overnight, some only do it for a couple of hours.

[DecomposingStar]

Leave it overnight. If your computer isn't a puddle of plastic in the morning, then you're stable... (exaggeration).

Alternately there is Prime95 and CPU Burn-in that you can test OC's on. Prime 95 being almost identical, and CPU burn-in (i'm not sure what it does) heats your Proccy up a hell of a lot.

[chuckskull]

To test stability use the blend test like suggested. The other tests usually produce a slightly hotter CPU but aren't a good indication of stability.

[GSte]

Just ran blend for 2 hours with no probs on:

E6600@3.4GHz with 1.375vCore, 1.53vMCH

Tried it at 3.6gig with 1.4 vCore and 1.53vMCH but fails almost instantly.....

Been messing around with it for the last few days, and it seems that if it's going to fail it does so within the first few minutes.... anybody else have an opinion on this?

[DecomposingStar]

Just ran blend for 2 hours with no probs on:

E6600@3.4GHz with 1.375vCore, 1.53vMCH

Tried it at 3.6gig with 1.4 vCore and 1.53vMCH but fails almost instantly.....

Been messing around with it for the last few days, and it seems that if it's going to fail it does so within the first few minutes.... anybody else have an opinion on this?

Sounds sensible, but the extra time won't hurt and gives a good indication of the heat your CPU will produce, and whether you components can take stress like that, that you would receive from say- playing a game for an hour or 2.

[Dark Horse]

New dual core OCCT just came out, worth giving that a go. It lasts 30 mins so if that passes then do your 12h prime. If it does then you know its not worth doing a 12h prime and to back down a little.