Is hardware cause of Display Driver failures?

[MPD]

Hi

Recently I'm getting GPU driver crashes when gaming (Civ VI particularly - running both DX11 or DX12) - this used to run very stably.

It seems to be when the GPU gets above around 80c, I've tried to set custom fan curves and can hear the fans really kick in - this delays the problem, but always still hit some magic threshold and the display driver dies and kills the game

I've reseated GPU, cleaned all fans (air duster) - updated display drivers to latest and nothing seems to be helping. Its odd that this problem has crept in and I'm wondering if it shows a failing h/w component (maybe PSU or GPU). Unfortunately this GPU used a non standard PCB layout and I don't believe I can retrofit a water-cooling kit

Rig details (an original Overclockers bundle from Sept 15)
Intel Core i7-6700K @ 4.6GHz

HS-031-CS Corsair Hydro H80i GT High Performance Liquid CPU Cooler
TeamGroup Elite 16GB (4x4GB) DDR4 PC4-19200C16 2400MHz​

MSI Z170A Krait Gaming Intel Z170 (Socket 1151) DDR4 ATX
SuperFlower Leadex Platinum 850W Fully Modular "80 Plus Platinum" Power Supply - Black
NZXT Noctis 450 White Full Tower Chassis
Asus GeForce GTX 980Ti Matrix GTX980TI-6GD5

Thanks in advance

Mark

 

[shankly1985]

Not always.
I have had more display driver crashing from a badly done overclock on the GPU.

 

[MPD]

Thanks for the reply... but I'm not overclocking the GPU. In fact I've tried underclocking it, and like the custom fan curve it delays but doesn't solve the issue.

CPU is overlocked (by overclockers at base, when I bought the rig)

 

[DmitryD]

Getting NVIDIA GPU driver crashes with recent update too, grrr

 

[mof]

Have you tried any benchmarks? Does it happen in all your games?

 

[MPD]

Thanks for replies... interesting on the driver issue. I might try to regress to a previous driver.

On the benchmark I've just successfully run a furmark - on the ASUS default "gaming mode" GPU setup. All seems normal - idling at 57c, quickly rises to 73c in 3mins - then slowly wanders through the 80s capping at 91c after 41mins. It then held steady for a further hour until I aborted the test.

So this isn't pointing toward a h/w fault.

I've now rechecked game files for corruption, done a complete uninstall of nvidia drivers (using driver fusion) - then a clean install of older driver (417.01). Still the same problem....

 

[FlyingScotsman]

I would stay clear of furmark and instead do a loop of say heaven/valley or superposition and see if there are any crashes.

 

[Dempsey22]

I had numerous per day playing any game. Thats with a Vega64.

I tried fresh install, loads of different drivers, switching pcie slots and so on.

None of them resolved my driver display errors. I gave up on it and they have mpagically stopped.

 

[MPD]

Done a couple of cycles of heaven and everything is stable. It seems to be a Civ VI problem - so its the old "update everything and keep dropping the graphics settings" solution.

thanks for your help all

 

[eeii]

I would try putting on v-sync so your card is not running at a silly refresh rate when it does not need to. I would also clean any fan fiiters if you need to as well.

 

[eeii]

If you are still having problems after this I would use DDU to remove the driver and re-install. it could be time to replace the thermal paste on the GPU. Depending on the cooler decent quality thermal paste can make a big difference compared with the muck they normally use. I've used hydro naught and its rather good. The creme de la creme is thermal grizzly conductonaut. But be very careful as it does not work with aluminum and causes a chemical reaction.