Signs of a dead GPU

[gamerguy319]

Ok, I just got my GTX 660FTW SIG2 video card back from the RMA and it was working normally for a couple hours and then started to artifact and crash repeatedly. It stopped finishing tests in 3DMark and Tomb Raider and Minecraft started crashing. So, I talked to an associate who said it may or may not be the GPU. Do you all have any suggestions on what the problem may be? The Power Supply is checking out and I have run MemTest and it checked out. (specs in signature)

 

[gcatalin]

Can you try the card on another computer ? If it acts the same on another computer then it's busted, if not, it may be another component. But this would be the first step.

 

[Nico641]

Odd behaviour of the cooling fans are usually an indication of something being faulty. Are they running on full throttle despite the strain? Did you update/reinstall your video driver?

 

[gamerguy319]

well I may have gotten some coil whine (pitch was constantly changing) on the fans but ill check and see. I tested 3 drivers: 314.22, 320.49, and the newest beta driver with the same result

 

[Nico641]

well I may have gotten some coil whine (pitch was constantly changing) on the fans but ill check and see. I tested 3 drivers: 314.22, 320.49, and the newest beta driver with the same result

If your fans are indeed faulty it could be possible your card is overheating and getting instable as a result. Did you monitor the heat of the card?

 

[gamerguy319]

yes, it got no higher then 71c, which if it is like most video cards is completely normal?

 

[gamerguy319]

and sadly I do not have another computer to test it on

 

[Nico641]

That is indeed normal.

Did you try running a benchmark with GPU monitoring to try to figure out if it's strain that's causing the failure, and if that's the case what amount of strain is doing it?

 

[gamerguy319]

no I did not. what would be a good monitoring software

 

[gamerguy319]

ok so I did the benchmark with CPUID HWMonitor and GPU-Z open and it got to graphics test 2 on cloud gate but it does not seem to crash because of a high load or temps. it just gave a bunch of purple squares on my screen with a completely messed up image and then exited out of the benchmark

 

[Nico641]

Yep, it's busted. Those guys at RMA are clearly professionals, they didn't fix your card. Hopefully you got guarantee on your repairs.

 

[gamerguy319]

Now heres the kicker... Can a bad pci slot give you delayed artifacting like that?

 

[Nico641]

Considering PCI slots are the main way of receiving and sending data for a GPU that would most likely be the case. Did you spot anything?