Question PC crashes and the fans get loud ?

[Summie]

Hello,
About a month ago this issue first started and has been occurring more frequently. When playing games, sometimes my monitor will go all black, and my GPU fans will get extremely loud. I'm still able to hear audio but if I'm in a discord call people wont hear me and the only way to fix it is to reboot. So far I've monitored my temps and nothing has been out of the ordinary, and I've also ran some stress tests which didn't cause a crash. At this point it's occurring a few times a night and I'm not really sure how to go forward with figuring out the problem.

Specs:
CPU - Intel Core i9-23900KF
GPU - NVIDIA GeForce RTX 3080 10GB
RAM - Corsair Dominator 32GB (16GB x 2) DDR5 5600MHz
PSU - 1000 Watt NZXT C1000 80 PLUS Gold
MOBO - ASUS PRIME Z690-P

I measured my GPU hotspot when it last crashed and it runs about 20C higher than the GPU temp but it was only about 95C when it happened so I'm not sure if that was the cause.

 

[stonecarver]

Two things Is the GPU dust free? If you blow it out don't let the fans spin!

Open the side panel and leave off and just as a test put a box fan blowing into the PC.

Than use computer like usual, do all of your regular habits. Go play those games that crash the system.

Later we will know if it's still crashing while gaming or the fan keeps it from crashing or not.

No crash = heat problem

Still crashing than we dig deeper.

 

[Summie]

Two things Is the GPU dust free? If you blow it out don't let the fans spin!

Open the side panel and leave off and just as a test put a box fan blowing into the PC.

Than use computer like usual, do all of your regular habits. Go play those games that crash the system.

Later we will know if it's still crashing while gaming or the fan keeps it from crashing or not.

No crash = heat problem

Still crashing than we dig deeper.

My GPU wasn't really dusty at all so I didn't have to worry about that. I took the glass off and put two handheld-type fans pointed right at the GPU and was able to play with no issues last night but then I crashed within 5 minutes of playing a game this morning.

 

[Misgar]

I measured my GPU hotspot when it last crashed and it runs about 20C higher than the GPU temp but it was only about 95C when it happened so I'm not sure if that was the cause.

That implies your GPU temperature is only +75°C which is below the +80°C limit recommended in this web site (at the end).

https://blinksandbuttons.net/how-hot-is-too-hot-for-rtx-2080/

They mention throttling at 88 to 89°C which could mean your hotspot will be up around 108 -109°C.

The RTX 2080 has a thermal threshold of around 88-89 degrees Celsius, beyond which it can start to experience thermal throttling and potential performance issues.

I've run my RTX 3060 at full power (170W) for many hours during video renders and it was stable, but only after Topaz fixed a bug in their software. Prior to that, when Topaz Video AI was still flaky, I set a GPU power limit of 95% in MSI Afterburner to stop the computer hanging.

I suggest trying a 95% power limit on your 2080 to see if solves the problem. It's not an ideal fix, but it'll help your investigation. Better to lose a few FPS but gain a stable system.

 

[Summie]

That implies your GPU temperature is only +75°C which is below the +80°C limit recommended in this web site (at the end).

https://blinksandbuttons.net/how-hot-is-too-hot-for-rtx-2080/

They mention throttling at 88 to 89°C which could mean your hotspot will be up around 108 -109°C.

The RTX 2080 has a thermal threshold of around 88-89 degrees Celsius, beyond which it can start to experience thermal throttling and potential performance issues.

I've run my RTX 3060 at full power (170W) for many hours during video renders and it was stable, but only after Topaz fixed a bug in their software. Prior to that, when Topaz Video AI was still flaky, I set a GPU power limit of 95% in MSI Afterburner to stop the computer hanging.

I suggest trying a 95% power limit on your 2080 to see if solves the problem. It's not an ideal fix, but it'll help your investigation. Better to lose a few FPS but gain a stable system.

Ah I accidentally put 2080 instead of 3080, my bad. Correct though, I'm only reaching regular GPU temps of about 75~ before I crash. If that could still be the issue with a 3080 though, I'll try to figure out how to power limit it.

 

[Misgar]

Temperature is an indicator the GPU is working hard, but +75°C shouldn't be enough to cause freezes/reboots.

I'd be more inclined to suspect software problems. My system with an RTX 3060 crashed regularly in Topaz Video AI back in early 2023, until Topaz fixed the bugs.

Similarly, games developers may release software before they've ironed out all the bugs. It's cheaper to let the end user run Quality Assurance.

Since you're running a 3080, I recommend visiting the NVidia web site and download the most up-to-date Gaming Driver. I use their Studio Driver which is designed for stability, but probably isn't ideal for you.

After drivers, I'd look carefully at the PCIe power lead connections to your card and make sure they're plugged in firmly. Check the PSU end too. Regardless of 12VHPWR problems, even "old school" 8-way connectors have been known to overheat on rare occasions.

If the problem persists on all games, I'd fit a cheap SATA SSD and install a fresh copy of Windows (after unplugging all other drives). It's vaguely possible that something has "broken" in your OS or programs.

Then I'd try another GPU, another PSU, another CPU and another mobo. Since this is not a practical solution for many people, do you or a friend have a (similar) PC you can drop the RTX 3080 into?