Hi all,
Hoping for a little bit of guidance in troubleshooting what I suspect to be a dodgy GPU.
Over the last 4-5 months I have been experiencing seemingly random crashes while gaming. About half the time I get a black screen for a few seconds before the computer restarts. The event viewer simply states "The previous system shutdown at --:--:--  on ‎--/--/‎---- was unexpected." The other case is where only the application crashes and I get an error message along the lines of "graphics runtime error" where the event viewer says "Display driver nvlddmkm stopped responding and has successfully recovered.". 
Temperatures appear to be normal and generally hover around the 60-75c mark when under load, and they don't appear to spike when crashes occur according to Speedfan temperature logs. This is the same for all other components. Sometimes I can go 8 hours straight without a hitch and a sometimes it'll crash 2 minutes into a session. I generally use my computer to play Destiny 2, but I have experienced the same problems in the Halo collection too.
 
OS: Windows 10 Education 64-bit
MOBO: Gigabyte Z170X-Gaming 7
CPU: Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz (8 CPUs), ~4.0GHz
Memory: Mushkin Redline Ridgeback G2 16GB
Storage: Mushkin Atlas Vital MKNSSDAV500GB-D8 M.2 2280 500GB
GPU: EVGA GeForce GTX980TI Superclocked+ ACX2.0+ 6GB
PSU: EVGA SuperNova 750 G2

Happy to provide any additional info that may help. Many thanks for your assistance in advance!
Cheers

Yikes... Sounds like something I would only want to attempt as a last resort. Is there anything else you could suggest to try and narrow down the fault before I take a shot in the dark on something like that?
Appreciate your response.

Thanks, I'll give that a go. At this point is it safe to assume that it's the GPU that's the problem?

I'm running 980 Ti's in SLI and I get this crash to unless I down clock as far as I can, -90MHz on core and -201MHz on memory.

I have noted that several elite video cards from that period were caught up in the change of solder to a lead free variant
 
The affected hardware galore including hard disks, motherboards and peripherals generally.
 
It took time to figure out how to make solders that were more reliable and not short out or fail over accelerated testing etc
 
 

veganfanatic
I have noted that several elite video cards from that period were caught up in the change of solder to a lead free variant
 
The affected hardware galore including hard disks, motherboards and peripherals generally.
 
It took time to figure out how to make solders that were more reliable and not short out or fail over accelerated testing etc
 
 

Wasn't the RoHS Directive for led free solder implemented in 06~09,thereabouts, for products coming in to the EU? I wouldn't have thought it would have affected a card made in 15, surely?

Puddles.

veganfanatic
I have noted that several elite video cards from that period were caught up in the change of solder to a lead free variant
 
The affected hardware galore including hard disks, motherboards and peripherals generally.
 
It took time to figure out how to make solders that were more reliable and not short out or fail over accelerated testing etc
 
 

Wasn't the RoHS Directive for led free solder implemented in 06~09,thereabouts, for products coming in to the EU? I wouldn't have thought it would have affected a card made in 15, surely?

EU adoption was clear back in 2003. There were already various formulations in the works years prior.

bob16314
Sounds like solder cracking somewhere on the PCB may have occurred..That was not uncommon with cards of that era and can cause issues such as artifacting, freezing, restarts, BSOD, more.

If all else fails, you can disassemble and bake only the PCB (aka the 'oven trick') to reflow any solder cracking that might exist, then hope for the best.

If you bake it, do it only with the GPU chip side up and not down.

Baking may only fix it for a short time, or a long time, or not at all if there's a component failure.

Disassemble the card..Preheat oven to 200C..Place the PCB with the GPU side up on a cookie sheet/baking pan..Bake for 10 - 15 minutes..Lots of other videos on YouTube about how to do it.

+1 
Seen my fair share of failing/failed cards and many have shown that sort of behavior. I've had great success with the oven trick, and may be worth a shot if you're not really wanting to spend money on a new GPU. I've had baked cards last day's while some lasted years after. 

bob16314
Sounds like solder cracking somewhere on the PCB may have occurred..That was not uncommon with cards of that era and can cause issues such as artifacting, freezing, restarts, BSOD, more.

If all else fails, you can disassemble and only the PCB (aka the 'oven trick') to reflow any solder cracking that might exist, then hope for the best.

If you bake it, do it only with the GPU chip side up and not down.

Baking may only fix it for a short time, or a long time, or not at all if there's a component failure.

Disassemble the card..Preheat oven to 200C..Place the PCB with the GPU side up on a cookie sheet/baking pan..Bake for 10 - 15 minutes..Lots of other videos on YouTube about how to do it.

Wow, thanks!  I have a GTX 980 Ti and was running into the same issue as OP and tried the bake idea mainly because I'm on the verge of getting a new card (I hope) and it WORKED!  At least in the short term my BSOD and hard resets seem to be gone.
 
Do you think that down clocking might help limp it along a little farther?  If so, I don't have much experience with it, would Precision or Afterburner be better to start with?

bob16314
 
If you use Precision, you should use for that older card.

Thanks!