Nvlddmkm \Device\Video3 Error occurred on GPUID: 900

No problem man.

i had this problem for MONTHS now with every effing game i played. i literally stopped playing destiny because i could not finish a single session without multiple random crashes. i had it with diablo 4, then ac6 and now it started with starfield. Since i did what you recommened i played without a single (!) crash for multiple days now. thanks for that.

here comes my question though: this solutiion seems to combat the symptoms (crashing) but i´m still baffled about the root cause. Is it an insuffient PSU? a faulty GPU? Hardware incompatability? Any further input from your side would be greatly appreciated.

Also sorry for my english i´m ESL.

EDIT: Disabling hardware-accelerated GPU scheduling seems to have solved the problem. In the last 3 days the problem has not reoccurred (before it sometimes occurred several times per evening).

Hello everyone,

I've been encountering a peculiar issue related to the nvlddmkm error, particularly when using Discord, and wanted to share my experience and observations in hopes of finding a resolution or aiding others with similar issues.

Preliminary Observations:

Prior to the driver becoming unresponsive and crashing with the nvlddmkm error (as per the Windows event log), one of two things typically happen:

1. Discord unexpectedly restarts without any apparent reason.
2. My own camera feed is not displayed to other participants in the call. When trying to restart discord to solve the camera problem, the crash occours.

Interestingly, this issue seems to be exclusive to Discord. I've run various games (e.g., CS:GO/CS2, Battlefield 3/4/2042, Cyberpunk 2077, Payday 3, MS Flight Simulator, Euro Truck Simulator 2, AmongUs, Squad, ...) and benchmarks (3DMark, PCMark, Furmark) without encountering this error.

Hardware and Software Troubleshooting:

• I've swapped out my RTX3080 FE with an older RTX3060 from MSI, which did not resolve the issue, prompting me to revert to the RTX3080 FE.
• Attempts to resolve the issue by reinstalling Windows and the drivers have been unsuccessful.
• The only component, that stayed the same is my Processor (Ryzen 5 5600X).

Recent Developments:

• The problem had not occurred for a while, but it's worth noting that since the last occurrence, I upgraded my Graphics Card to an RTX3080 TI from Gigabyte. The problem did not reoccur until today.
• A friend of mine has been experiencing the exact same problem for some time, which adds an intriguing layer to this issue. His Graphics Card is an RTX3080 Ti from EVGA

Detailed Incident:

1. In a Discord call, my camera was not being displayed to my friend.
2. Upon restarting Discord on his end (without a graphics driver crash), my camera feed was still not visible.
3. When I attempted to restart my Discord, my graphics driver first stopped responding, then crashed, necessitating a hard reset of my PC.

Doubts Regarding Hardware Issues:

I'm inclined to believe that this is not a hardware issue like mentioned by mkxy since it has occurred with four different cards from various vendors/partners.

Next Steps:

We have disabled the Hardware-Accelerated GPU Scheduling now. Due to the spontaneous nature of the problem we cannot say if it helped or not, but I will follow up in a couple days to share if this changed helped or not.

I'm curious if anyone else has encountered similar problems (regarding the specificity to discord) or has insights into potential solutions. Any feedback or shared experiences would be greatly appreciated. Thank you for taking the time to read through this detailed post.

@ventbyte you seem to be confusing a lot of things. I never said it is caused by a physical hardware defect inside the gpu silicon or vram. I said it is caused by a power instability. either by using OC, undervolt, broken PSU, broken mainboard, or messed up drivers which interfere with power supply to the card. I said how you mostly can fix it, by reducing power limit in MSI AB by 20% and also setting a global FPS limit inside Nvidia control panel between 60 and 120 fps for example. the crash will happen with either 100% GPU or 100% VRAM load, lowering the power profile and capping FPS will buffer critical power instabilities. you didnt seem to have tested the workaround, because you didnt say anything about it. Applications which cause a memory leak inside the VRAM can also cause this issue btw.

Started having this issue with my GTX1080 after recent Windows Updates/Driver Update, not sure with which specific one.

• Changed cables.
• Changed GPU. (I had a spare for testing.)
• Normally use stock clocks, tried undervolting.
• Changed displays.
• Reinstalled affected games.
• Completely cleaned out the drivers and reinstalled.

In the end, this isn't hardware instability. This appears to be related to the driver or Windows. Turning off GPU Scheduling and/or using the Debug option under Nvidia seems to have bypassed the issue, but it isn't a solution. The solution would be for the vendors to correct the new issue.

If you're asking yourself "Why am I being told to undervolt/underclock?" and the reason this might help some users is that it'll simply make your card slower and less likely to encounter the problem. It isn't a solution, it's just a way to avoid the problem.

In the end, your card isn't unstable and there's been plenty of feedback here expressing plenty of examples against that.