I work on computers for a living, but this has me stumped. Having a random restarting, apparently due to memory dumps in Windows (according to event logs and dmp files), and I am completely unable to isolate anything as the culprit.
Whenever I'm playing a game (ie: CS:S, LOTRO, etc) I get random mem dumps. The machine freezes for a second or two, clicks, then just blanks out and reboots. It may not do it for 2 or 3 hours, then all of the sudden hiccup. Usually after the first time, it does it again shortly after. This happens repeatedly.
I've run a crapload of hardware diagnostics multiple times and have had zero failures. Suspected RAM first, ran Memtest86 a few times, including about 10hrs straight yesterday (32 passes) and no problems. Even brought home a new pair of 512's to test and it still acted up. So that pretty much rules the RAM out.
I ran a couple video RAM tests that all passed, and 3DMark 05 -and- 06 both ran through the basic test fine. I had the issue with both my old PSU and my new 430W Thermaltake. HDD tests out fine for read/write/stress. CPU stress tests fine. It's definitely not a heat issue. I have no frickin clue whats left.
I'm beginning to suspect the motherboard. Does this seem like the only remaining possibility? If so, I'm about to just upgrade mobo/CPU/RAM anyway. Any suggestions or advice appreciated. TIA
Well it's apparently not the audio. The BSOD is too quick to read, and it still restarts. I'm pretty sure it's been narrowed down to CPU/mobo or running out of RAM. I could understand if LOTRO is hogging the RAM with only 1 GB, but certainly not CS:S.
Last night for example, it ran fine for about 3 hours before it finally froze and restarted. After the first time, it did it a couple more times closer together.
When not in 3D mode, or just idle, the machine will run all day long just fine. Will run diagnostics and memtest literally all day without a problem.
So many things can cause bsod... since it passes all of the tests known to man and it is solid when not in 3d (and if you want to assume video is fine, which it probably is) then the only thing left is the mobo. I would have gone for heat first, then a PSU which it sounds like you have done (your new PSU is still kinda skinny for your setup).
You have my permission to drop a half-dime and get a new MoBo, cpu, and MORE memory. If you can wait a little - go for a quad-core @ $266(your cost will most likely be around $325). If nothing else, get 2x2G memory now, it's _really_ cheap...
Personally, I am waiting for bearlake, ddr3, and 45nm but if you gotta go - (like my ol momma used to say) life's a bitch (, you marry one,) and then you die...
I would check for a heat problem somewhere. 3 main areas are CPU, RAM and Video card. Especially the video card will give the random restarts, like when you're playing a game. I know with my card, the default fan speed was way too low so I had to manually set it and the restarts stopped.
Being that everything easily swappable has been swapped, that narrows it down to CPU/board. It's only a minor overclock, and all the voltages are stock except CPU which is at 1.55.
I now bumped the chipset up one notch from 1.2 to 1.3 and the RAM from 2.7 to 2.8 to see how it runs; maybe an undervoltage issue due to OC'ing the system. Just trying to make it efficient enough to last me a while 'til I can afford to upgrade.