1080ti blank/blue/green screens crashing or white squares crashing/rebooting

MIDIBoss

Weaksauce
Joined
Aug 5, 2017
Messages
125
Hello all,

I have a 1080ti. Bought it new in 2017. I'm literally a month away from RMA warranty no longer being good on it. This year because of Covid I started gaming a lot more. I got a 4K LG OLED TV about a year ago. I always gamed in 4K since then, it's awesome. In March I started having some white squares all over the screen on Warzone. I posted a picture of what's happening below, not my image but it is the same as what's happening to me. Then it would cutoff the whole machine and crash the whole computer. Only way to get it back on would be hard reset holding power switch, turn off computer, and reboot.

Now I am still having that issue and/or depending on the game, I also have random crashes to a black, blue, green, or blank screen which results in the same way: No sound and no video, must hard restart the computer. I cannot finish a 10-minute race of Assetto Corsa Competizione without the later problem occurring, for example.

I built this PC myself. 3 years and no problems until now. Here are specs:

i7 8700k
1080ti GPU
32gb G.Skill RAM
Corsair SF 600 watt PSU
Noctua NH9 cooler
lots of case fans
there's a Firewire card installed with an nvme m.2 adapter
nvme m.2 Samsung 960 PRO SSD
Ncase M1 V5 case
I think that's it

Do you suspect GPU or PSU failure?

Here's a clue; When I use EVGA Precision software to downclock or underclock by -5% on all available options,it does not crash. EVGA customer support suspects PSU problems.

Also I get Kernel 41 errors in Event Viewer system log, power related.

I already tried DDU deletion/cleaning of Nvidia software/drivers and reinstalling Nvidia latest driver manually. Windows up to date. Don't know what else to check.

I want to figure this out!!!! Cyberpunk 2077 is coming out in 2 months! I need to be ready!!!!! :)

Thank you.

MIDIBoss
 

Attachments

  • 0106f833-026b-4924-829f-167862cc1710.jpg
    0106f833-026b-4924-829f-167862cc1710.jpg
    577.5 KB · Views: 0
That's a definite gpu failure.

Thank you for your reply. One thing I want to point out: When I use MSI Afterburner or EVGA Precision to *LIMIT* the ultimate power of the card to 95% or less, the problems go away. Still think GPU failure? Thank you.

MIDIBoss
 
Mine was doing something similar and I cooked it. Problem went away, but she was no 1080ti and no bloody warranty. If you still can, perhaps rma? Cars would still be worth something after fix

Thanks for your reply! Do you mean it was the video card or the PSU that failed? I can still RMA the video card for another month! The PSU is under warranty for an additional 4 years.
 
Thank you for your reply. One thing I want to point out: When I use MSI Afterburner or EVGA Precision to *LIMIT* the ultimate power of the card to 95% or less, the problems go away. Still think GPU failure? Thank you.

MIDIBoss
Yep... It's failing. Eventually you will get the memory corruption squares even with lower power and clocks. Get it rma'd while you can :).
 
Hello all,

I have a 1080ti. Bought it new in 2017. I'm literally a month away from RMA warranty no longer being good on it. This year because of Covid I started gaming a lot more. I got a 4K LG OLED TV about a year ago. I always gamed in 4K since then, it's awesome. In March I started having some white squares all over the screen on Warzone. I posted a picture of what's happening below, not my image but it is the same as what's happening to me. Then it would cutoff the whole machine and crash the whole computer. Only way to get it back on would be hard reset holding power switch, turn off computer, and reboot.

Now I am still having that issue and/or depending on the game, I also have random crashes to a black, blue, green, or blank screen which results in the same way: No sound and no video, must hard restart the computer. I cannot finish a 10-minute race of Assetto Corsa Competizione without the later problem occurring, for example.

I built this PC myself. 3 years and no problems until now. Here are specs:

i7 8700k
1080ti GPU
32gb G.Skill RAM
Corsair SF 600 watt PSU
Noctua NH9 cooler
lots of case fans
there's a Firewire card installed with an nvme m.2 adapter
nvme m.2 Samsung 960 PRO SSD
Ncase M1 V5 case
I think that's it

Do you suspect GPU or PSU failure?

Here's a clue; When I use EVGA Precision software to downclock or underclock by -5% on all available options,it does not crash. EVGA customer support suspects PSU problems.

Also I get Kernel 41 errors in Event Viewer system log, power related.

I already tried DDU deletion/cleaning of Nvidia software/drivers and reinstalling Nvidia latest driver manually. Windows up to date. Don't know what else to check.

I want to figure this out!!!! Cyberpunk 2077 is coming out in 2 months! I need to be ready!!!!! :)

Thank you.

MIDIBoss
Do you have a known-good power supply you can test with, or a known-good system you can test it in? That power supply is perhaps marginal for that setup, but it should be OK, assuming it's healthy. More importantly, what you're seeing here isn't really characteristic of a failed power supply. Nevertheless, if able, you should test with a known good power supply to confirm this.

I suspect a BGA or GPU package failure. If you have a month or two left in the warranty, I say RMA it now. Things get a lot sketchier if you have to start thinking about repairing the card.
 
My assumption is also a vram issue due to the squares.
Its 'possible' the PSU is causing it, gfx ram failing before anything else on the gfx card - due to a power problem, so I would test the PSU first.
 
I assume your card is an EVGA right? If so RMA it, I had a similar issue with with an EVGA 1080ti, they were super helpful and even allow cross ship if you put a deposit cc on file so literally no down time. Depending of stock you may get offered a side grade of card maybe 2xxx series even. Even if another vendor still worth it as you dont want to be out a card but might be a pita to get done.
 
Do you have a known-good power supply you can test with, or a known-good system you can test it in? That power supply is perhaps marginal for that setup, but it should be OK, assuming it's healthy. More importantly, what you're seeing here isn't really characteristic of a failed power supply. Nevertheless, if able, you should test with a known good power supply to confirm this.

I suspect a BGA or GPU package failure. If you have a month or two left in the warranty, I say RMA it now. Things get a lot sketchier if you have to start thinking about repairing the card.

PSU test couldn't hurt but I agree that you should start the RMA now before time runs out.

I would also leave cooking the card to reflow solder as an absolute last resort - I once tried this on a few failing cards with these results: no change on one, fixed one for about a month, outright killed another card.
 
Hey all,

Thank you very much for your collective input. I requested an RMA and I will report back when the replacement gets here! To me it didn't make total sense to just stare at the 12V voltage reading without the machine being under load, but the tech at EVGA suggested I go into BIOS and look at the 12V rail meter for a few minutes to see if the voltage fluctuated. It never did, stayed right at 12V the whole time, so we went ahead with the crosship RMA. I'll report back when I get the replacement card and let you guys know if that fixed it! Thanks again. Great to have the help. 1st time I ever dealt with this.

Cheers
MIDIBoss
 
Cross ship option is nice. Didn't have that luck with another vendor, just gave up as i was likely getting the same card back and just got a new one.
 
Back
Top