ASUS Maximus Formula SE

[jonny_ftm]

funny, at 3.825 im OCCT beta stable

but i cant run large prime for more then 20 minutes without an automatic bluescreen/restart

@safan80. thanks!

Yep, hence why it's better to test many programs for benching. I'll try to give you some situations showing that necessity, and limitations of every program

OCCT in RAM mode proved to be the more stressing and revealing errors on NB-RAM combo. Do remember, that even an error code 4 (CPU) can be due to a NB or RAM error (vNB, vDDR, RAM timings, STRAP...). If the NB and RAM return erronous values to the CPU, than OCCT and any other program will have a round off error check and read it as a CPU error, so be careful interpreting the results... (Sine I changed my RAM to Hyperx 9600 and used a 4:5 multiplier to get 974MHz on them, I need 1.55v on my NB instead of 1.45 for 1:1, otherwise OCCT in RAM mode will return me an error code 4). The more you OC your RAM, the more you need to volt your NB too, even if the FSB is unchanged

Also, OCCT in CPU mode should be as effective as Prime95 small FFT in testing CPU stability. On the other side, to check cooling limits of the CPU, small FFT is better as it generates more heat. Do remember, that no program will generate as much heat as small FFT (if we exclude the crazy intel TAT).

Finally, a F@H for some hours should be a necessary check, as it is a real program that can much stress the CPU-NB-RAM all together.

As a Prime95 Blend test uses more RAM than OCCT, it can reveal more hardware defects in RAM than OCCT, like Memtest would do. Also, Blend test will heat the CPU more than OCCT in RAM mode, thus unveiling an unsufficient CPU cooling. But definetely, OCCT in RAM mode, will stress your RAM an NB much much more than Prime Blend.

Hope my explanations show you why you need many programs to test your OC stability.

The final test, will be every day use to see if you have reboots, BSOD or random strange program or memory access violation errors