Game Causing Whole PC to Crash After Several Minutes of Play

Technical Support
1

Since the update on 11/19/24, Diablo IV causes my entire computer to shut down after playing for about 4 minutes. This has happened every time I’ve tried to play the game since that day.

There is no error ID/crash message b/c the monitor just goes black, the computer stops working, and the cooling fans whir up very high speed. I have monitored CPU and GPU temps just before the crash. They were ~72 and ~45 respectively.
I have run the “Scan and Repair” function on Diablo IV from the Blizzard launcher, and it found no issues.

I’m using Windows 10; NVIDIA RTX 3090 with the most updated drivers (version 32.0.15.6614); 12th gen intel core i9-12900k 3200 Mhz, 16 cores; 64 gb RAM.
I also tried rolling my drivers back to a version from 8/14/24 (version 32.0.15.6094), at which time I was having no issues with the game.
I have not experienced crashes or other issues with any other program.

When I look for reports in Windows Event Viewer, so far as I can tell, these crashes do not generate a report. By which I mean, when I got to Event Viewer, select Application, and search for “diablo” in all types of events, there are no reports for the past several days. This is despite having the game crash several times over the past 2-3 days. The only errors it shows for this evening are with Bonjour Service, and they did not occur while the game was running or just prior to the crash when I tried to run D4 tonight.

I have tried Blizzard’s “Get Help” page, but it just it’s not caused by their product, and it’s an issue with my machine. Maybe, but when literally no other game causes this? Kinda hard to believe.

Does anyone have recommendations on how to fix or further troubleshoot this issue?
Thanks.

2

I would run a test to stress the GPU and make sure there isn’t an issue there.

You can download FurMark here:

https://geeks3d.com/furmark/downloads/

You can use version 1 or 2, doesn’t matter much. Download, install and just run the default test.

Let it run for 10-20 minutes or longer.

Make sure you are on a FRESH RESTART of your system, and don’t run any other tasks.

It will take at least 5 minutes for the GPU to heat up and show the loading, which is why I recommend letting it run as long as you can.

Now.

If while you are running this test, you see visual artifacts in the image that is being rendered, or the system crashes, that means you have an issue, potentially with the hardware. That being heat, or failing card or failing power supply.

But as I said, you need to run the test for a good length of time.

You can also use this tool from passmark:

https://www.passmark.com/products/performancetest/download.php

This will install a trial version, but it has a more comprehensive battery of tests.

Again, just run the default set tests, you do not need to change the test parameters for either of these.

1 Like
3

Thanks very much for the advice!

I ran FurMark 2.0, and it triggered the same crash as Diablo IV at ~3:15. Just before the crash stats were:

FPS ~130
GPU UTIL 98%
GPU Temp 72.5C (“Hotspot” was 92.0 C)
GPU CLK 1170 MHz
GPU V 0.712 V
GPU PWR 324 W (Power draw seemed to be ~100% on average.)
GPU Fan 1230 RPM

Interestingly, I ran it again with V-sync on. All the stats seemed to hit “steady state” at about 8 min. I kept it running for 20 min. At that point I extended the display to my 2nd monitor and started streaming a show, since I do that a lot when playing games.

With v-sync on while running FurMark for 20 min, then running FurMark & streaming a show on a 2nd monitor for another ~6 min., the stats at ~26 min were:
FPS ~60
GPU UTIL 38%
GPU Temp 68C (Hotspot 84.4 C)
GPU CLK 1740 MHz
GPU V 0.893
GPU PWR 256W
GPU FAN 1035 RPM

I don’t know what to make of that information, but it was the first program I’ve been able to run that recreates the problem.
Any input would be much appreciated.

4

Well, this is a major indication.

Your hotspot is waaay above what I would feel is normal.

I ran Furmark 2 last night. Hottest I could get that same part was 69c! The GPU was 62c.

You didn’t get as hot in that second test.

Time to shut it down and blow out the dust. First and easiest fix. Don’t use a vacuum, and don’t let the fans spin too fast while blowing out the card’s heatsink and fans.

Second fix is to address any airflow issues with the case. Make sure the video card has an unobstructed airflow around it. Make sure there is plenty of air movement through the case. Make sure you have as many fans pulling air out as you are putting air in, and hopefully so it pulls air in from one end so it goes out the other.

Thirdly you can use Afterburner to create a custom fan curve to more aggressively cool the card when in use. I do use this on my card. I don’t use it to overclock or change voltages or anything else. I simply use it to adjust the fan spin up thresholds so its more aggressive at cooling the card.

1 Like
5

Cool. I’ll give it a shot.

Thanks again!

6

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.