Question 1080ti crashes constantly in VR games, unless slightly underclocked

[Celauron]

GPU: Asus Strix 1080ti 11gb
CPU: Intel 10100KF
PSU: Seasonic 750W

During very high loads inside VR games graphics driver/GPU appears to become unstable, and games freeze/crash as well as run laggy and choppy. They produce standard errors such as:

LowLevelFatalError [File:Unknown] [Line: 198] Unreal Engine is exiting due to D3D device being lost. (Error: 0x887A0006 - 'HUNG')

Which implies graphics driver or device issue. While it happens, the gpu temps do not exceed 70C.

Nothing similar happens when running stress test in Furmark, or the exact same games on highest settings non-vr, that said I do not have any non VR games that would stress the GPU as much as VR games. There are no graphical artifacts or bsods even when playing VR games - they just hang on a jittering image, and either produce an error, or get stuck until I manually close them through task manager. In case of one particular game (No Mans Sky) - I can hear the game playing behind and my character being able to open inventory, while the screen simply hangs on a jittery loading image.

Nothing like that happens when using 970 GTX spare card. And it seems that nothing like that happened on the previous PC this GPU was used in, even inside VR games.

Reducing the GPU core and memory clock by 100MHz using Msi Afterburner seems to fix the issue and for some reason produce less or even no lag at all. Severely lowering the resolution in VR and/or in-game settings also solves the issue.

I did clean installation of drivers after removing them with DDU in safe mode. Tried increasing fan rotation. Tried various different graphics settings.

Any suggestions on why this could be happening? I understand why lower resolution in VR helmet would fix the issue (older gpu/cpu not handling the game) but why would underclocking work as well?

 

[Roland Of Gilead]

Hey there,

Please list your full PC specs, including mobo and PSU.

 

[Celauron]

Hey there,

Please list your full PC specs, including mobo and PSU.

GPU: Asus Strix 1080ti 11gb
CPU: Intel 10100KF
PSU: Seasonic 750W
Motherboard: Asus Strix b460F
RAM: Corsair 32GB 4x8 2400 using XMP2 profile to up to 2666Mhz, nothing else is overclocked

Just in case it was overheating after all - I tried to up the fan even more and used HWInfo to monitor temps - even hotspot temp never went above 86C. I also declocked with -25 core and -90 memory mhz and slightly lowered in-game settings from highest to high, this resulted in 2 hours of gameplay but then the same issue occurred. Yesterday, I played more than 3 hours of half life Alyx, albeit declocked 100 core and memory both, and on lowest settings and did not experience any crashes. However, prior to declocking - it would freeze the game and either crash or force me to close through task manager.

At this point I am just trying to understand if it's simply the case of low specs, either CPU or GPU not handling the games, or GPU being faulty. If it is faulty - I don't understand why would it work in Furmark and on Ultra settings in normal games, and why doesn't it produce bsods or artifacts. I have also increased TdrDelay timing to 10 seconds, but I don't think it mattered - the crash still happened sudden and randomly.

If I increase the settings in-game or increase VR resolution - I either can't load the games at all, or crash upon start on even in-game menus, though that doesn't help to narrow it down.

I have also tried OCCT benchmark with the highest VRAM usage possible - also no issues occurred, at least after 15 minutes.

I have tried running 3dmark and it resulted in immediate GPU crash.

Event Viewer produces this:

The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Error occurred on GPUID: 100

The message resource is present but the message was not found in the message table

As well as

Display driver nvlddmkm stopped responding and has successfully recovered.

I can also see previous errors from VR games:

\Device\Video3
Reset TDR occurred on GPUID:100

\Device\Video3
Resetting TDR occurred on GPUID:100

\Device\000000e3
Error occurred on GPUID: 100

Lastly - declocking -130 both core and memory allowed to finish 3dmark without any issues.