Prime 95 Critical error

[redguard117]

Hey guys, i built a rig around a year ago. Unfortunately I was too lazy to run Prime95 back then, i only ran 3dmark. I've noticed during this time that the computer freezes at times - it doesn't just freeze, whenever i make a rapid movement with my mouse the speakers produce this repeated thumping/beeping sound (cant really describe it, its like nothing i've ever heard). Anyway i decided to run Prime95, ran it for 6 hours without a problem. Tried again for 8 hours, no probs.

Anyway, i decided to run it one more time today after several months. I start it, leave my pc on, and leave the house for 7 hours. I come back to see that

[Jan 5 07:37] FATAL ERROR: Rouding was 0.5, expected less than 0.4
[Jan 5 07:37] Hardware failure detected, consult stress.txt file
[Jan 5 07:37] Torture Test ran 7 minutes - 1 errors, 0 warnings.
[Jan 5 07:37] Worker stopped.

So apparently i failed prime95 after 7 minutes... Although i'd never failed before. My ram and cpu are running at stock settings. However, i did notice that this crashing that i mentioned before usually occurred when i was running my ddr3 at its rated speed of 1600. I bumped it down to 1066 and the crashing reduced, although it did not complete stop. I don't know if this is related but I also get a bsod whenever i run nvidia ntune to change fan settings, and nvlddmkm at times chooses to stop working. I tried overclocking a long time ago for a very modest 30 mhz boost and my computer did not boot.


My specs are

Intel C2Q @ 8.5 multiplier, 1333 fsb (stock settings)
EVGA 790i Ultra
Corsair XMS3 DDR3 1600 (9 9 9 24 1.9v)
Coolmax 950 Watt PSU
MSI GTX 260
Tagan EL Diablo case
WD Caviar 640 Gb

My cpu fan is a cnps9700 which I always run at full speed, and my mobo and gpu fan's also always run fullspeed

Any suggestions, help would be greatly appreciated in figuring out why my rig is so unstable. BTW this rig is NOT connected to the internet so i highly doubt its malware, unless i infected my rig with a usb flashdrive or something.

 

[redguard117]

I noticed that i'm having the same exact trouble as this guy, and he has almost the same exact rig.
http://www.tomshardware.com/forum/260540-30-system-freezing-corsair-xms3-ddr3-1600#t1829674

 

[ukcal]

When I was OCing my e7300, it got to a point where I was getting that error. Personally, I solved this issue by raising the voltage to the CPU. Strange that such errors are occurring at stock speed, but consider a small voltage increase to see if that works.

 

[drums101]

+1 to Ukcal whenever I would stress test an OC and get that I would just raise the vcore a little bit and it would be fine. The wierd thing is that its happening on a stock running cpu...did you adjust the VID so it will run cooler or did u jus pop the chip and not mess with anything in the BIOS?

 

[redguard117]

i basically left the bios alone except for setting my memory settings, voltages and such. i tried to oc a while back but apparently my cpu is a dud because it refuses to boot even 30 mhz above the usual fsb.

Following your suggestion I set the voltage at 1.23, originally i was using Nvidia's "auto" settings but looking at what they put my memory at thats not a very good idea.

I just ran IntelBurnTest using the new voltage and it was perfectly stable. Perhaps its my memory? I posted a link above to a few other guys who were running almost exactly my same rig (same mobo, cpu, gpu, AND ram) and apparently the combination leads to the same crashes

well anyways thanks alot drums101 and Ukcal. Looks like I have an excuse to play some WaW since i need to ...um... make sure my rig isn't unstable. the problem with testing it though is that its just so d@mn unpredictable it can happen on an empty desktop, when i'm running prime95 and realtemp at the same time, or when i'm donig a research project.

still, any excuse to get a new rig is welcome