For the past few weeks I have been having this really annoying issue with the new Nvidia graphics drivers. It first started with the 197.16 drivers and continues with the 197.45 drivers. The problem is that whenever I play a game the Nvidia drivers crash along with the game and I need to restart my PC in order for them to recover. This used to be a once a day thing which occurred hours into gameplay, and It used to only happen on heavy games like Battlefield BC2 and Killing Floor. I thought I fixed it by OCing my graphics card to a 700clock speed (linked w/ shader). - After OCing i was able to play for a week without any annoying crashes. However, now It has morphed into an even worse problem in which now the drivers and the game crash almost immediately after executing a game (30 secs or so).
Is my card dead or is there something I am missing here?
I dont think its the GPU overheating, I have GPUz and evga precision running now to see temps during the crash and It is not going over 70c on World of Warcraft. and not going over 78c for Killing Floor.
I have a strong feeling it may be the PSU. When I first built my PC, the readers on Toms Hardware warned me about my PSU purchase.
Here are the specs:
Intel(R) Core(TM)2 Quad CPU Q6600 @ 3.2GHz OC, ~2.4GHz
NVIDIA GeForce 9800 GT
Windows 7™ Ultimate
Well, the most likely culprit in that list is your POS PSU. That's a very low-quality unit. It doesn't have active PFC, only has 72% efficiency, and puts out most of its output on the 3.3v and 5v rails. Modern systems need output on the 12v rail. A low quality PSU will cause all sorts of stability issues.
That's quality RAM, but have you performed an overnight run of Memtest86+ to test for RAM errors?
The idle temps look OK, but what really counts is the load temps. Have you watched the CPU temps while putting the CPU under 100% load for an extended period of time using something like Prime95?
I took a second look at SpeedFan while running Prime 95 and saw that at idle my fan was really only going around 400rpm (30c). while under load (55c) it was only going just under 900rpm. That sounds very low to me. I ran Prime 95 again to confirm this. It seems that my CPU fan does not scale with my CPU temps. I reached 60c and the fan went DOWN to 300rpm... my 4th and 1st cores failed the test within the first 4 minutes.
I dont know if this is how PSUs work exactially, but this is my grasp on what is happening. With the power output increasing in the CPU (while running prime 95), a decrease of power follows in other parts of the computer such as the CPU fan. Following this logic, this would explain the low fain speed warning i get at startup and the game crashes whether its my graphics drivers failing or the game just not responding.
You're either having an overheating problem or your overclock isn't stable. Does the system crash with the CPU at stock speed?
It seems like your HSF isn't working correctly. You could disable any auto fan settings in the BIOS and make the fan on the heatsink run at full speed or you could pick up a new cooler that completely bypasses the stupid push-pins and screws on.
aha! I found a major problem! Arctic Cooler Freezer 7 Pro was completely burnt out! I am not sure if this is directly related to my graphics driver/game crashing issue, but I tried turning on my PC one last time after cleaning off the CPU fan and noticed that I never even started up. I confirmed that the the Fan was broken by plugging it in to my old PC which resulted in the fan not even budging! Im glad I found this problem I am gonna have a look at the fan you mentioned in the link above. Thanks for your long term assistance shortstuff!