Dead gpu?

Associate
Joined
7 Nov 2013
Posts
24
Last night while playing left 4 dead 2, my game flickered and eventually went pure grey and locked up.

So i shut the thing off, and checked behind the case / interior.
I had been cleaning the interior of the case , so the cables were near my feet and not organized. The psu power cable was very loosely inserted, and had probably been like that for a few weeks without me realizing.

So i rebooted again and tried playing l4d2 to see if the issue would happen again. And it did.
This time when i boot, the windows boot screen was pixelated and stopped loading midway, frozen.

Eventually windows 10 booted into repair mode after failing to start twice. Once the repair was done, my background image was all messed up and files were damaged / things not working as they should(Avira AV said it was corrupt and was disabled). I checked the error viewer and had loads of nvldmkm nvidia errors.

I took the GPU out and reseated it, and reinstalled windows 10. Once it restarted for a windows 10 creation update 1709, the boot windows 10 boot screen was green and pixelated, worse than before. Frozen once again.

https://i.imgur.com/DoLcjDM.jpg

I'm now running on a old HD 6670 radeon card instead and reinstalled windows 10, with no issues to report, i managed to play left 4 dead 2 on low for a good 2 hours without issue and have been using it all day.

I guess my question is, how do i guarantee this is my gpu thats broken? This radeon card does not require 8 + 6 pin power connector to function, so i'm thinking its either the PSU is not delivering the proper power requirement for the 780, or the GPU really is dead.

Ironically, i was looking to rebuild this week anyway prior to this happening(arriving soonish once the OEM 8700k chips are in stock), due to my machine being 5 - 6 years old now, it was a prebuilt from overclockers back in 2011 .

I had a look at the sticky, it says to check voltages with HWMonitor, only the 5V one looks to be over the maximum value, im not sure if thats a problem or not. Of course, these volts could show differently once the 780 is plugged back in right?, but its highly unstable to really check to make sure.

https://i.imgur.com/LKjMkwV.jpg

Although i've got a new rig on the way. This one aint that bad, i was hoping to sell it along with the 780 secondhand. But this has really thrown the spanner into that idea.

Specs of rig in question:

Titan Spinosaur" Intel Core i5 2500K 3.30GHz @ 4.60GHz DDR3 Sandybridge System
8gb 1600 ram
Gigabyte GeForce GTX 780 WindForce 3x (Upgrade i bought for the rig back in december 2013 from overclockers)
Psu: OCZ ZS Series 750W 80 Plus Bronze Power Supply
 
Last edited:
Worth monitoring the GPU temperature (which the above will do), or downloading Afterburner etc.

Could be the GPU is overheating, and may just need a fresh application of thermal paste.
 
I had afterburner installed, the gpu never went over 70c before, but i havent been watching it lately to be certain. I'll go back and check the temps to be sure though.
 
Alright, plugged it back in, installed the drivers, noticed a flicker on the desktop Downloaded heaven and ran it.

As soon as i launched benchmark my screen started to have a flicker rave (HWMonitor turned black around the UI also), until eventually the entire screen turns grey and doesn't recover. PC Shutdown this time though, then it detected no GPU.

I'm back on the other GPU, report center is filled with windows hardware errors around when i started the benchmark.
"A problem with your hardware caused Windows to stop working correctly."
LiveKernelEvent

Took some pictures of the benchmark, you can see the temps didn't reach above 50c
https://imgur.com/a/Wy6Wy
 
Last edited:
Back
Top Bottom