0
Completed

General ingame crashyness, blue screen even, nvwgf2umx.dll

AtomicStryker 1 year ago • updated 1 year ago 4

Windows 10 x64 (not on Creators Update, otherwise up to date)

Intel i5-6600K, Geforce 1080 Ti, 16 GB Ram

WFTO Version: 'v1.6.3' via Steam


It doesn't crash on startup or level load, it crashes after 30 Minutes of gameplay or more. But it crashes almost every time. It even caused Blue Screen of Death, wow i have not seen those in years.

I am playing other games such as Overwatch or Doom just fine, so it's not some kind of hardware failure.


Attached the latest crash files and the output log.


It looks like you or the Unity engine are doing something the Nvidia driver does not like at all.

I was running 382.52 at the time - i am now updating to the most up to date ones.

But WFTO is an older engine? I dunno.


output_log.zip2017-08-05_202804.zip

2017-08-05_202804.zip


Game Version:
Steam Public
Platform:
Windows
Pending Customer

Hi Atmoic Stryker,


Sorry to hear about your issues with the game, it's really frustrating when you're faced with crashes like this. I've taken a quick look through your output logs and it seems that the stacktrace is consistently pointing to the following:


0x00007FFF8FE8EA22 (nvwgf2umx) NVAPI_Thunk
0x00007FFF901AF547 (nvwgf2umx) NVAPI_Thunk
0x00007FFF8F6C4BEC (nvwgf2umx) OpenAdapter12
0x00007FFF8F86FD5F (nvwgf2umx) OpenAdapter12
0x00007FFF8F86FC06 (nvwgf2umx) OpenAdapter12
0x00007FFF8F86817D (nvwgf2umx) OpenAdapter12
0x00007FFF8FF3FBAA (nvwgf2umx) NVAPI_Thunk
0x00007FFF905A4E41 (nvwgf2umx) NVAPI_Thunk
0x00007FFFC9478364 (KERNEL32) BaseThreadInitThunk


Unfortunately I'm not the most qualified person to be looking at these but I am the only chap around at 1:20AM on a Sunday so I will naturally be seeking the opinions of the more technical personnel on our team. However my understanding is that this indicates the problem is not occurring within our code but instead within the Nvidia drivers. If my knowledge is right this is actually below unity and "closer to the metal" than anything we or unity touch, the fact that you suffered a blue screen also supports this as generally we would never be able to do anything to cause a complete kernel failure.


Of course it is also peculiar that this is only occurring only in WFTO when your experiences is games like Overwatch and Doom are fine. My first guess would be that it could be a D3D call is falling over at some point and causing the crash. 


A Quick google of nvwgf2umx turns up a number of threads with similar issues, seemingly random crashes in a number of games (such as GTA5). 


My advice for the moment would be to do what you're already doing, update your graphics drivers and see if it helps the issue. If that doesn't help you might need to try something else, from reading the output log I see you have a peculiar virtual resolution, are you using a spanned multi-monitor system and spanning the game across both displays? If so try disabling one display temporarily and seeing if that helps to resolve your issue.


I found this Reddit thread for a similar issue in GTA5 which might have some extra steps: https://www.reddit.com/r/GrandTheftAutoV_PC/comments/36qwab/suffering_from_frequent_crashes_here_is_a_guide/ in particular I'd look at ensuring that your redists aren't broken.


Please let me know how you get on in the meantime I'll bring this to the attention of our QA chappy and the programming team, but it's possible you may only receive a response during next week. Good luck!


Cheers,


Lee


Alright, i have investigated aswell and several issues came to light:


- i had turned off VSync as i generally do to improve responsiveness and FPS

- WFTO happens to be one of those games that have no internal FPS lock and also can burn all 4 CPU cores at 100%

- it turns out, that did overwhelm my CPU cooling over time


- i have adjusted my CPU cooler to provide more cooling power

- i have re-enabled VSync as it limits the games FPS and CPU usage (90 FPS is fine)


The problem could be fixed, as it was indeed hardware failure on my part. But only time will tell, won't it ^^

I do apologize if it was indeed user error.


It is strange that the crash would point to nvidia runtimes when the CPU was overheating (GPU was balmy), but then again, when the CPU starts miscalculating anything can happen.

Completed

Great news, we'll mark the ticket as complete for now but please let us know if anything changes and we'll do our best to help.


With regards to the stability of your CPU it certainly can cause anything to turn up a fault though it's indeed surprising to see it presenting in NV Driver stuff. Out of curiosity have you OCd your CPU if it's getting really pushed perhaps the OC is somewhat unstable when the CPU is under extreme load.

Well i had loaded a profile in UEFI that allowed the CPU to use the "TurboClock" feature so yes. It did basically overclock.


That being said i would not claim the PC wasn't stable if i didn't actually check with Prime95 and Furmark (which led to the overheat discovery above). I don't do this periodically, and apparently the CPU cooler had loosened at some point.