Prime95 testing question

[sanityvoid]

Hi guys,

I'm having an issue with a stable overclock. I can't get a stable overclock on my older CPU after 3 hours. I have a AMD X2 4200+

I was thinking that in using Prime95 that I didn't use the right test for testing the different segments of the OC.

For example using Small FFT instead of using Large FFT for the different component tests.

I think I used small FFT for each OC section, but just in case could someone refresh my memory on what to use for each test? I did reread Prime95 readme but I got no answers from that. Or maybe I just missed it.

use Small, Large, Blend and Custom for

CPU/FSB: ? (pretty sure I used small here)
HTT: ? (used small here)
Mem: ? (same as above)
HT: ? (same as above)

Any clarification would be great, thanks!!!

 

[richardscott]

use small fft always you have to test mem with memtest. you cant test htt or ht, if they are unstable they will make small fft unstable.

 

[jsc]

Large and small fft run from the CPU L2 cache - minimum memory I/O.

Testing my Q6600, I used 3 instances of large ftt and one instance of blend.

Large fft's reputedly maximize the load on the CPU cores.

 

[olddman]

I ran the large FFT on my Q6600 after I got the machine together. It does a 100% load test on all cores. It's running now after I OC'ed to 3.0GHz. Beem running for 3.5 hrs now and so far so good.

 

[iluvgillgill]

this is how it works small ffts works the CPU to its max and mobo and ram stay normal.blend test(aka large ffts) use alot of memory.this is to test overall system stability as ram is most stressed but because data pass from CPU to ram through the motherboard(AMD-HTT,Intel-FSB) so the whole system is stressed.