1080ti crashing AFTER stress tests

[Hocknose]

I've had a ASUS Rog Strix 1080ti (non OC) for about 3.5 years. I've recently built a new PC and it has been completely stable for 3 months. In the last few weeks I started to get crashes where my screens would go black and the GPU fans would go to 100%. Windows was still running in the background but the PC would need to be hard reset.

I've performed a number of things to try and fix the issue:

• Replaced PSU

• Reinstalled older drivers via DDU

• Reinstalled Windows 10

• Cleaned and reapplied thermal paste to the GPU (GPU temps were not an issue but thought I would try it)

So after all these didn't help I took my base unit to a friends and we swapped out our 1080ti's and my GPU did the same crash in his build. This narrowed the issue down to it definitely being the GPU itself.

We did a number of stress tests on the GPU and what we found was that the GPU tended to crash most often just after the Stress test completed. We noticed this in:

• Furmark

• 3DMark

• OCCT

Its a strange issue because it doesn't crash at all at the height of the stress test but just after. Has anyone come across this issue before?

[DoctorNick]

What does event viewer say? Does it bluescreen?

[Hocknose]

No never BSOD. Doesn't always show anything in event viewer but when it does I get:

 

Hardware Error

 

Problem Event Name:    LiveKernelEvent
Code:    117
Parameter 1:    ffffa70f17ec8460
Parameter 2:    fffff80363c83ee8
Parameter 3:    0
Parameter 4:    4dc
OS version:    10_0_19042
Service Pack:    0_0
Product:    256_1
OS Version:    10.0.19042.2.0.0.256.48
Locale ID:    2057
 

 

[Mark Kaine]

That's a weird one, but it does seem like a power delivery issue to me, just on the card itself not psu or motherboard. 

 

Can you physically check the pcb of the card for any burnt signs, blown capacitors etc?

 

 

[Hocknose]

1 hour ago, Mark Kaine said:

That's a weird one, but it does seem like a power delivery issue to me, just on the card itself not psu or motherboard. 

 

Can you physically check the pcb of the card for any burnt signs, blown capacitors etc?

 

 

I think that will have to be my next thing to test, will try it tonight

[Hocknose]

I can't see anything that looks blown to me but I'm no expert. Took the card apart and gave it a good clean but still having issues

[Maticks]

PSU's can do all sorts of things, if you can borrow another from a friend you can rule it out if you don't want to buy a replacement.

 

[Hocknose]

I've already swapped out the PSU on my own PC and I got the same failure in a friends PC. So I've tried it with 3 different PSU's now