How much burn-in test is enough?

ziphnor

Weaksauce
Joined
Jan 16, 2005
Messages
65
Hi,

In OC'ing my new E6600 ive been using ORTHOS to test stability. I was testing 391x9 and since ORTHOS had run a few hours on small FFTs i thought everything would be okay. I left it on while i went to work and when i came back it had failed after 6h 50m !

I then very slightly increased the voltage ( so that speedfan says 1.50V before vdroop(but the BIOS setting is 1.5750, it always gives less voltage than i choose, apparently thats just how this mobo (P5B-E) acts ). Additionally i lowered the FSB to 389, bringing me to 3.5Ghz. I have now run ORTHOS with small FFTs for 11 hours and 30 minutes (and its still running).

How long is enough to be 'sure' it wont fail at all?
 
orthos stresses the cpu in ways that games never will. I've never run orthos/prime longer than 3-4 hours, and I'm perfectly stable. However, I can't boast "24 hour prime stable".

For a loss of 18 mhz, and a couple degrees higher temps, whats the difference? Go with the more stable one.
 
Naturally i will go with the stable one, i would just like an OC that is just as stable as if i had done no OC at all!

I ran Large FFTs and it failed after ~6 hours so the other test was not sufficient.

I have upped the vCore a tiny bit and when i get the time i will run a long large FFTs test.
 
ziphnor said:
Naturally i will go with the stable one, i would just like an OC that is just as stable as if i had done no OC at all!

I ran Large FFTs and it failed after ~6 hours so the other test was not sufficient.

I have upped the vCore a tiny bit and when i get the time i will run a long large FFTs test.
Doesnt the "LARGE" ftt test the memory? So if it failed, one could assume that a memory issue is the problem.
 
thrawn86 said:
orthos stresses the cpu in ways that games never will. I've never run orthos/prime longer than 3-4 hours, and I'm perfectly stable. However, I can't boast "24 hour prime stable".

For a loss of 18 mhz, and a couple degrees higher temps, whats the difference? Go with the more stable one.

Now I dissagree with that. What if Intel sell you processor they claimed stock at 2ghz and it turns out to be unstable after orthos 6 hours @ stock speed? And you have to reduce the speed to reach better stability ?

Stability is stability. It's should be 100% stable with all the tools we have.
I agree with go with the more stable one. So find the max setting where the system is stable say 18hrs+ then reduce the clockspeed a bit.
 
newls1 said:
Doesnt the "LARGE" ftt test the memory? So if it failed, one could assume that a memory issue is the problem.

Well my memory is running at their rated speed(well actually slower than that) and what i got was a rounding error. I think its the blend mode that really stress memory.
 
Back
Top