Faulty Graphics Card?

[rhuynh]

Greetings,

I'm running into an issue with my computer where during gaming sessions, the screen would go black and I would hear the fans of the GPU go loud and I'm forced to hard reset. If I re-seat the card, reapply thermal paste, or use a different driver version, I sometimes get the error that the display driver has stopped responding and has recovered instead of the black screen. After recovery though, I experience graphics tearing and I have to restart anyways.

My main computer specs are:

Intel i5 3570k @ 4.2 GHz
Asrock z77 extreme4
Seasonic X-Series X-850
Corsair Vengeance Blue 8 GB (2 x 4 GB) 1600mHz DDR3
MSI Radeon HD7870, R7870 Twin Frozr 2GD5/OC @ v13.4

I have ran memtest overnight with no errors, and then Prime95 and Heaven simultaneously overnight with no errors. The crashes only occurs when I have a game open, and it has never been consistent. Once, I was able to make it a full day without it crashing but after I turned it on and started playing the next day, it would crash.

When the crashes occur, the temperature, fan speeds, and GPU load are usually constant so I do not believe it is an overheating problem. The temperature of the GPU according to GPU-Z averages around 54c during my typical load, and 72c during a Furmark or Heaven stress test. The fan speeds are set to auto and vary between 30-40%. The load during the average crash is typically around 60%.

I have tried multiple things like replacing my PSU, reinstalling and using different drivers making sure I use Driver Sweeper in safe mode to completely wipe out the old driver before installing the new drivers, and underclocking the GPU clock and memory speeds by 100Mhz. I've also made sure that the drivers/bios for my motherboard is up to date. I've also tried reinstalling DirectX.

Unfortunately, I do not have another computer to test components on one by one.

It really confuses me that stress tests can't recreate the problem, but whenever I try to play a game that doesn't run my system at 100% load, I run into issues. Does that probably mean that the problem isn't hardware related, but how the entire system works together?

Any suggestions would be much appreciated.

 

[nanaki333]

had the same problem (with latest BIOS and drivers). running drivers 12.4 i think it is will fix the problem. i can't remember for sure, but to get rid of it and use the latest drivers on my wife's computer (had a 7970), had to update the onboard intel iGPU drivers. for some reason that's what was causing her 7970 to do the exact same thing as yours is.

 

[rhuynh]

Interesting. Would installing my motherboard's vga driver help then? I left out that driver because I thought that AMD's driver would override it.

 

[nanaki333]

there's a good chance that will fix your problem, yes. i skipped that driver as well actually. i googled for a looooooong time and finally found someone that said to try that. i had a post very similar to yours. it was driving me nuts with 7970 in xfire with an sb-e system, and later with my wife's sb (then upgraded to ivy) system.

can't guarantee it will work, but i would give it a try.

 

[rhuynh]

I found the option to just disable the onboard iGPU. After that, I tried to apply the VGA driver and it said it couldn't find the supported hardware which is probably a good thing.

I wasn't able to test it out under normal conditions since I used the evening to do more stress tests, but that solution would explain one of the reasons for why the system works under heavy load but not under normal conditions. I didn't know that the system could jump between using dedicated graphics and the onboard graphics if that is what's happening.

 

[nanaki333]

oh nice! you know, i searched that bios for a long ass time and couldn't find where to completely disable the iGP. either way, hope it works out for you with that thing out of the picture.

 

[rhuynh]

For my motherboard, it was under advanced -> North bridge configuration -> IGPU Multi-Monitor (Enabled/Disabled)