Please help! nvlddmkm.sys TDR error

hdiffut

Member
Joined
Jun 28, 2024
Posts
8
Specs and dmp attached. displays lose signal then reboot a few seconds later. All bios, drivers and firmware up to date. Happens 4 or 5 times after first booting up. Eventually it stays up and I can game for hours without issue. also happens after waking up from sleep mode. Tried DDU, physically reseating the card, full system wipe. Starting to feel like my GPU is dying. I saw a similar thread on here and disabled C-state in the BIOS. didn't help. Not sure how to proceed. You guys have helped me out before. Hopefully you can save me from having to RMA this 3080.
 

Attachments

Apologies. Attached is my file collection app report.

OS - win11 pro x64. license purchased
self built desktop built in 2022
Ryzen 9 5900x
mobo - ASUS TUF Gaming x570-PRO WIFI (bios 5013)
RAM - 2 16gb dimms in slots 1 and 2 DDR4
PSU - EVGA 1000w gold
GPU - EVGA RTX 3080 FTW
no OC, no VPN, no disk imaging tools
 

Attachments

All but one of the dumps show the Windows Timeout Detection and Recovery feature (TDR) attempting to reset the graphics card and graphics driver because of a detected graphic hang. (The other is a DPC watchdog timeout, probably because of a long running nvlddmkm.sys ISR). The resets don't clear the problem and we get the 0x116 bugcheck. Often this is down to a bad driver (yours looks to be up to date) or a bad graphics card. The dumps don't tell us much else however, nvlddmkm.sys is the only third-party driver referenced in the dumps.

I can see from your logs that there will also be a bunch of dumps in the C:\Windows\LiveKernelReports\Watchdog folder. We don't need these, but (from your logs) these are dumps written because of graphics failures from which Windows was able to recover - these probably resulted in a crash to the desktop rather than a BSOD.

The best way to check whether this is a driver or card problem is as follows...
  1. Download the most recent three driver versions from the Nvidia website. Also download DDU.
  2. Use DDU to remove the current driver.
  3. Install the latest driver.
  4. If it BSODs then use DDU to uninstall the driver.
  5. Install the next back-level version of the driver.
  6. Repeat the above process until you either find a stable driver (in which case it was a bad/incompatible driver).
  7. If it BSODs on all of the most recent three drivers then it's most likely to be a card failure.
If it looks like the card then try removing and reseating it and ensure that the additional power cable is fully home at both ends.

The acid test will be if you can beg or borrow another graphics card to test. If you do this make sure to use DDU first to remove drivers for the old card.
 
Back
Top