One step at the time is the best way to isolate the the limiting factor. The obvious choices are going to be either the Mobo, RAM, or CPU. One other possibilty that comes to mind is that the MOBO doesnt like that particular brand of RAM. Also you can easily get instability with an ify PSU. Is the RAM on the MOBO's QVL? To isolate the problem, begin by setting the RAM back to loose timings and a slow frequency. I am guessing that the multiplier can be reduced on your CPU. If so then reduce it lower than stock. Now find your Mobos max FSB. Since you have taken the CPU and RAM down to below stock settings this should allow you to OC the MOBO higher than you may have done previously. Bump the FSB until it becomes unstable and then either move to the next component or add a touch of voltage to the chipset. Write down your max FSB. Now run your stability test with your RAM at stock or below settings. If all is well then start OCIng the RAM and testing its stability in between increases.
Ive got a feeling that the Mobo doesnt like your RAM for some reason.
The RAM has been running fine for over 2 months without incident, the RAM seems perfectly compatable with the board.
Anyway, I ran mem test on each stick seperatly and they both passed, I let the program run 2 passes.
Ideas? - I hope the CPU isn't the problem, as I sanded the heatspreader for laughs and it got rid of the printed details of the CPU, thus voided warrenty!
If the RAM is compatible then your going to need to isolate each component and find their OC ranges. How much of an increase in FSB have you achieved with this system?
Um, I don't want to overclock, my sh|t aint running fine @ stock.
Right, so far - Im mem test stable when testing one stick at a time.
If testing both sticks I get errors.
I am Prime stable with both sticks for at least 35 tests (stopped it because I got bored), which is wierd, becuase test night it would fail straight away
Is this the Socket754 Neo board? If it is, I recall that RAM timings should be set to 2T to run two DIMMs.
It's quite possible it would work for a long time set to 1T. It was probably marginal at that point, and something (like overclocking? ) slowly pushed it over the edge.
I'm not surprised it would pass Prime95 and not MemTest. MemTest puts RAM through some pretty esoteric usage patterns that Prime95 might not even use. Plus, you'd be surprised how often RAM can lose a bit here and there and still not cause any obvious problems (at least not right away). I recall one pretty difficult memory test--one that memtest doesn't do--is to stuff the RAM with data and leave it holding that data for a few days or even weeks, then come back and check data integrity. A lot of otherwise stable RAM (especially non-ECC RAM) won't pass that test!
<i>"Intel's ICH6R SouthBridge, now featuring RAID -1"
"RAID-minus-one?"
"Yeah. You have two hard drives, neither of which can actually boot."</i>
The mem test errors have decreased, when I first started to notice the problem I ran mem test and got 3xxxx errors, last time I check I only got 10-50 errors.
Ok when you run the sticks individually you dont get any memtest errors, but together you do get probs rite? You tried clocking the RAM to 333Mhz? You might have knackered your CPU's memory controller when you overclocked.
could try up'in the vcore of the cpu to what it was when u were oc'ing and see it it passes for longer? And does your system stay stable on the desktop at stock? Or does it stuff up (kick u out of apps and freeze?)
I started noticing problems as windows would randomly restart...but, the odd thing is, Windows is not running fine, and has been for about 10 hours.
The errors in mem test seem to becoming less and less as I stated earlier (from thousands, down to around 10). I haven't changed any settings since I first posted. Only thing I've really done is take out clean and replace the DIMMs.
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.