Question GPU Crash Help

[Hammiches]

I've been having an issue with my computer where playing games on high or even medium settings will crash my GPU.

The crash: When the crash occurs the two monitors plugged into the GPU go black and say that there is no signal. My third monitor, plugged into my motherboard, will flicker on and off as my computer tries to make it into my primary screen (the monitor is in portrait mode and is set to be portrait, but after a crash it will want to be horizontal).
I've usually been able to get my computer to return to normal by simply restarting it, but occasionally I've had to wait a couple hours, or, more quickly, open BIOS on bootup. I don't have to change anything in BIOS, I just have to open it.

As far as I can tell my graphics drivers are crashing when my GPU has to use VRAM past a certain point.

Solutions I've tried so far: cleaning my computer, reinstalling my BIOS, and a hard reinstall of display drivers using DDU.
Nothing has fixed it.

The hardware:

GPU: Zotac GeForce RTX 3060TI

 

[Hammiches]

I've been able to mostly avoid crashes by turning down graphics settings well below what a 3060TI should be able to handle but it'll still crash at specific points in the games.

 

[Gaidax]

That's not much info here to go on with. Do list your full configuration including CPU/Motherboard/PSU/Overclocking if any and any other info to work with, so there will be something to work with.

 

[Hammiches]

That's not much info here to go on with. Do list your full configuration including CPU/Motherboard/PSU/Overclocking if any and any other info to work with, so there will be something to work with.

Happily, I had written something longer but it wasn't letting me post it because it was too long.

No overclocking whatsoever.

PSU: Corsair RM750x
CPU: Intel Core i7-12700k
CPU cooler: Corsair Water cooling (forgive me I forget the model)
Motherboard: ASUS Prime Z690-P Wifi D4
Ram: 32 GB DDR4
SSD/HDD: Samsung SSD 980 Pro

 

[Hammiches]

Yeah, let's not download random executable files from account created today with 3 posts in a forum.

Thankfully, I'm smart enough to avoid that.

 

[Gaidax]

Happily, I had written something longer but it wasn't letting me post it because it was too long.

No overclocking whatsoever.

PSU: Corsair RM750x
CPU: Intel Core i7-12700k
CPU cooler: Corsair Water cooling (forgive me I forget the model)
Motherboard: ASUS Prime Z690-P Wifi D4
Ram: 32 GB DDR4
SSD/HDD: Samsung SSD 980 Pro

TBH hate these kinds of issues, it can be like 5 different things and it's all trial and error, so don't expect here anything more than further suggestion to try out.

Do you have a warranty on that GPU? Can you test it with another PC? Do you connect it with separate PCIE power cables? How long do you have this GPU and when the issue started?

Can you track its temperature to see maybe that is the issue?

 

[Hammiches]

TBH hate these kinds of issues, it can be like 5 different things and it's all trial and error, so don't expect here anything more than further suggestion to try out.

Do you have a warranty on that GPU? Can you test it with another PC? Do you connect it with separate PCIE power cables? How long do you have this GPU and when the issue started?

Can you track its temperature to see maybe that is the issue?

I greatly appreciate your help.
I would bet the warranty is expired, I built this computer almost exactly 3 years ago. I don't have another PC to test it with (and I wouldn't actually know how to even if I did). This issue started roughly 4 months ago.
The question about the PCIE cables is interesting, I'm not sure if I do. I'll have to open it back up to check. Would you clarify what you mean when you say separate PCIE cables? As in, I should have 2 cables running to the GPU?

I've tracked the temperature in the past in a game that had been crashing and it was within an acceptable range but I'll happily track it again. If you have a software you recommend that would record other data around the crash I'll run that too.

 

[Hammiches]

TBH hate these kinds of issues, it can be like 5 different things and it's all trial and error, so don't expect here anything more than further suggestion to try out.

Do you have a warranty on that GPU? Can you test it with another PC? Do you connect it with separate PCIE power cables? How long do you have this GPU and when the issue started?

Can you track its temperature to see maybe that is the issue?

I'd also like to add that I did get a crash code one time after a crash and I saw it associated with a "Hung GPU" online. Regrettably I didn't write down that code, but maybe that narrows things down.

 

[Gaidax]

1. In regard to the cables - ensure it's either of these setups, I imagine it is but hey weirder things happened.

2. For temps tracking MSI Afterburner or HWMonitor can give you the info. HWMonitor is pretty straightforward - just keep it running on that 3rd display and when crash happens check what were max temp readings for the GPU.

3. Try to see if the issue happens with only one monitor connected and no others connected at all including the one to the iGPU.

4. Downclock the CPU/VRAM a bit with MSI Afterburner and see what happens. You can also try undervolting it, but you need proper guide for that one. The purpose is to see if with lower clocks or power consumption it stabilizes.

That's just the ideas that are not straight RMA. I'd still check the warranty, maybe it's 5 years or maybe you're still within 3 years standard.

 

[Hammiches]

Temperature hit 74°C before the crash. GPU usage was at 99%. Is it possible that's the culprit?

 

[Gaidax]

Temperature hit 74°C before the crash. GPU usage was at 99%. Is it possible that's the culprit?

Those temps are fine. Usage is also fine.

Reduce GPU frequency by 10% in the afterburner and see what happens.

 

[Hammiches]

Those temps are fine. Usage is also fine.

Reduce GPU frequency by 10% in the afterburner and see what happens.

Unfortunately, it seems like this made it crash faster.

 

[Gaidax]

Unfortunately, it seems like this made it crash faster.

Try to undervolt it instead then, there are plenty videos in YouTube for it showing the process for your GPU specifically.

See what happens. If still this issue happens, try to use only one single monitor connected to your PC, with two others disconnected.

After that I'd run out of ideas, except for RMA if possible.

 

[Hammiches]

I resolved to do this tomorrow and started playing some WoW classic and now, for the first time ever, WoW classic is causing the crash. Monitoring on Afterburner it's hitting the same marks as something way more demanding. 100% GPU usage, 74°C. No way a game from 2004 should be causing so many problems. Do i have some virus in the background?

 

[Gaidax]

I resolved to do this tomorrow and started playing some WoW classic and now, for the first time ever, WoW classic is causing the crash. Monitoring on Afterburner it's hitting the same marks as something way more demanding. 100% GPU usage, 74°C. No way a game from 2004 should be causing so many problems. Do i have some virus in the background?

There is no virus, you simply have uncapped framerate and your GPU is pushing all it can, which CPU is enough for the job, that's why you see 100% usage. It does not really matter if game is from 2004 or 2024.

I really smell and RMA here, if possible. Even if undervolting would solve the issue, it would only solve it for just a while anyway.

But who knows maybe some people here would provide other ideas.

 

[beyondlogic]

I resolved to do this tomorrow and started playing some WoW classic and now, for the first time ever, WoW classic is causing the crash. Monitoring on Afterburner it's hitting the same marks as something way more demanding. 100% GPU usage, 74°C. No way a game from 2004 should be causing so many problems. Do i have some virus in the background?

what is your cpu temperature.

how is your aio mounted top or front of the pc case.

pc case name ?

how is cpu intensive more then gpu