Over the last couple of months I've had system crashes and "nvidia kernel has stopped responding and has recovered" errors while playing some games. Borderlands 2 and Guild Wars 2 seemed to cause the crashes more frequently, while League of Legends never crashed. Fast forward to a couple weeks ago and now the system is crashing while surfing the web, some images on the web are corrupt, and playing anything other than LoL isn't even possible, and even the new ARAM map is causing issues! It's time to fix this shit, so I begin diagnosing: -Full format and reinstall of my OS, plus install all the original drivers that came with all the components. System crashes while browsing the web, ok its not a software issue. -Memtest86 is run on one stick of ram, makes it through 3 passes before Memtest freezes. No errors, screen locked up. -Run it again on the other stick, makes it 4 passes before it locks up. Ok, isn't a RAM issue but something else. -Update BIOS, still crashes. -Then I notice something odd. While in my BIOS I'm watching the voltage for my CPU jumping around at nearly 10% fluctuation. 12v is going way high low, from 11.88 to 12.244. Ok, PSU is probably going out. This sucks because I had to RMA my first one, as the fan speed controller was DOA. -Pull out my old PSU from a previous system and plug it in. Bam, zero crashes for 48 hours playing any game I have. Go to the local parts store and get a new PSU and plug it in. That was 3 days ago and I had zero crashes in a 10 hour GW2 session. The next day I decide to update my drivers to the current version since I was using the drivers that came with the video card and they're 9 months old. Launch GW 2 and play for a bit. Crash. "Nvidia kernel has stopped responding and has recovered." Fuck. Revert back to the next revision back since there's a ton of people complaining about stability on the 320.18 drivers. 314 series installed, reboot, launch GW2. 15 minutes later it crashes. Now I'm installing the original drivers again after removing physx and the 314 series. Hoping for the best at this point! /TL;DR - Fuck system instability, Antec for their PSU, and Nvidia for their shitty drivers.
I had the same problems with a Geforce 570. Exact same thing. At first, it was ocasionnal, at the end, always crashing. I RMA'ed the card 3 times. There was something flaky in the card (RAM?) that was causing these crashes. As soon I relaced the card, everything ran smoothly. Any other vidcard to test? Each time, crashes were less frequent after a complete reinstall.
It's just so crazy that a complete reinstall of the OS can stop the crashes temporarily. Makes it feel like a software issue.
It makes me think that it's just another case of M$oft doing more unauthorized, unwanted and under the table updates that aren't playing nice with drivers.
When I changed the card completely, the problem was solved. The card was crashing on any OS I tried it. That's why I suspected a material problem. But we can't be sure at all!
