BSOD Caused by nvlddmkm.sys

I'm not sure where to put this, since most of my issues are non-BSOD issues, but I feel like they are related.

I just recently built a new computer, and besides a monitor I had to send for repairs, it was working fine until I upgraded the PSU, added a second graphics card with NVLink SLI, and added a Samsung 970 Pro. The issues started off with Windows thinking there was something wrong with the graphics drivers. Upon booting into Linux, it would state the NVLink was unlinked and only one graphics card would show up (the second one I put in). The issue continues to happen until I power off the computer completely, turn off and on the switch on the PSU, and restart the computer.

After having that issue a few times, I upgraded the BIOS and drivers on Windows, and while it was idling earlier today, I got a BSOD, namely SYSTEM_THREAD_EXCEPTION_NOT_HANDLED caused by nvlddmkm.sys. The computer hit 100% on collecting info and froze, forcing me to hit the reset button. I later tried checking for RAM issues via a test Windows provides, and that too froze. I reset the computer, checked BIOS settings, and upon saving and exiting, another freeze. I shut down the computer and did the same thing I did to "fix" the NVLink issue from before. Computer is working fine so far, but I'm certain I will continue to have issues.

Basically, I'm wondering what the likely cuplrit(s) are.

I ran the Beta log collector just in case you need it: Attachment 228064

Update: I just disabled C-States just in case that was causing some (if not all) of the problems; will report back to see if that clears some of the issues.

I have listed below a number of Eventlog entries from the last 2 days:

Code:

2019-03-19T14:27:54.158		Windows failed fast startup with error status 0xC00000D4.
2019-03-19T14:29:11.391		Force to reinstall graphics driver.
2019-03-20T19:31:22.452		Display driver nvlddmkm stopped responding and has successfully recovered.
2019-03-20T19:31:26.424		Display driver nvlddmkm stopped responding and has successfully recovered.
2019-03-20T19:44:12.106		Dump file creation failed due to error during dump creation.
2019-03-20T19:44:13.384		The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

There was also the following entry which will explain why you are not always getting the crash dump files:

Attachment 228128

Some things to do:

1. Turn off Fast Start-up
2. Consider putting the pagefile back to the C: drive while troubleshooting
3. try removing all the Nvidia stuff with DDU and then just reinstall the latest driver and PhysX

If you need more help on how to do the above just ask or look up in the tutorials.

Alright, I did the following:

1. Turned off Fast Start-up
2. Removed all the Nvidia stuff with DDU, and reinstalled the latest driver + PhysX
3. Moved the pagefile back to C: (not sure why or how it automatically set it to D:)

New crash: Attachment 228147

This time it looks like it's DPC_WATCHDOG_VIOLATION, with faulting driver nvlddmkm.sys.

On the bright side, my computer hasn't frozen from this, and the NVLink seems to be working still.

I disabled a couple of Razer programs that I think were causing the issue (probably not NVLink compatible), and so far, Windows seems stable. I'll update within a week if I don't have any more issues, and if there are none, mark this as solved.