OCCT error on memory test

[mpogr]

Hi all,

I have the following system:

eVGA 680i Mobo (rev. A1)
QX6700@3.3GHz, 1.475v
2x1GB OCZ XLC 9200@1100MHz, 2.1v
eVGA 8800GTX@stock, NB@1.3v

The system is water cooled, including NB and memory.

The system seems very stable at these settings and passed Prime95 with varous FTTs for many hours with no glitch. OCCT CPU and CPU&RAM test also passed for more than 4 hours.
Super PI 32M and WPRIME are also OK.

However, OCCT memory test always fails after ~25 minutes.

I tried everything, including running everything @stock, but the error still happens.

What should I do? Is my system really stable? Is there any known flaw in OCCT?

Thanks!

[strange|ife]

how many hours with prime? have you run orthos large fft's?

if all passes at least 8-10 hours, i would call it stable

is the system day to day stable, temps within reason? if so dont worry.

never tried occt though sorry, cant comment on software errors within it

[mpogr]

Tried large FFTs for more than 10 hours.

Temps are kindof high, reached 80C CPU and > 70C per core with Prime95 4 threads and small FFTs.

My point is the OCCT error is very predictable, seems to happen always at the same time and, most importantly, happens at stock and even lower settings.

I'll try to lower temps by introducing second water loop. Right now I have too much stuff in it:
CPU->NB->GPU and memory in parallel to all these. This probable kills water cooling efficiency.

[Bail_w]

Just run the OCCT 30 mins auto test twice or so. if you pass, your system is stable.