Hey Csema,
Thanks for feeding back, that's good to know about the XMP, and looks like it's already helped a few others.
Much appreciated!
B_M.
This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
I wanted to share how I resolved the frustrating nvlddmkm issue many people continue to experience.
TL;DR:
One of two DIMMs (RAM sticks) on my computer was faulty, causing errors in the GPU's ability to consistently access the graphics driver.
Removing the faulty DIMM solved the issue.
The issue:
Many users (myself included) experience a crash while playing games or using graphic-intensive applications.
This can manifest as a black screen, crash to desktop, unresponsive system or sometimes immediate system restart.
In Windows Event Viewer, this can show as a Warning from Source: Display, Event ID 4101:
"Display driver nvlddmkm stopped responding and successfully recovered."
It also appears in Unreal Engine crash reports as an Unreal Process crash:
"Unreal Engine is exiting due to D3D device being lost".
Issue appears in 3D Mark as a crash during benchmarking, with a benchmarking score of 0 and notification "an error has occurred".
Interestingly, Heaven Benchmark did not encounter this issue.
Attempted fixes:
Many people have posted solutions to this issue, and sometimes they work for some users, but never did for me.
These solutions include:
**I must stress, updating the system Registry can have negative repercussions on system stability if done incorrectly.
The solution that worked:
The Windows Memory Diagnostics Tool didn't find any issues with my RAM, but it got me thinking. If there was a problem with only one of the two DIMMs, could the tool determine this?
From some late-night YouTube watching, I learned that RAM is actually pretty darn important to the system, and also that RAM performance can degrade or even fail over time. Luckily, it's comparatively cheap to replace relative to the other computer components.
I removed one of the two DIMMs, and found the computer wouldn't boot when powered on.
I bought two new DIMMs to replace the old RAM, and hey-presto, 3-years of nvlddmkm problems were now resolved, and has never re-occurred 6-months later.
Summary:
If you are experiencing nvlddmkm error crashes, driver updates/software solutions may work for you.
If not, you may have faulty hardware, and this could be the RAM, like I experienced.
Test your RAM by removing one of your two DIMMs, and then swapping them and testing again.
Luckily, RAM one of the easiest and most affordable components to replace.
Hope this helps.
Ben.
Locked Question. This question was migrated from the Microsoft Support Community. You can vote on whether it's helpful, but you can't add comments or replies or follow the question.
Hey Csema,
Thanks for feeding back, that's good to know about the XMP, and looks like it's already helped a few others.
Much appreciated!
B_M.
Any news? Was your previous RAM Corsair?
Previous RAM was Corsair LPX, new ram kingston fury. Unfortunately still getting the issue.
Symptons:
"Display driver nvlddmkm stopped responding and successfully recovered."
"Event ID 0 from source nvlddmkm"
"Event ID 14 from source nvlddmkm"
"UCodeReset TDR occurred on GPUID"
"Lowlevelfatalerror...Unreal Engine is exiting due to D3D device being lost"
Have tried (didn't fix):
Increasing TDR registry setting (just hangs for longer before finally closing)
RAM
NVIDIA Driver (various old versions)
Only having a single monitor plugged in
Disabling rebar & 4g in mobo settings
Next I just noticed that "turn off harddrives" was set to 20mins in my power plan, it's not likely that the game didn't talk to the ssd in 20mins and it went to sleep, but it's worth a shot I guess.
Update (also didn't fix):
HDD Sleep setting in powercfg disabled
Prefer max performance in nvidia GPU
Stopping "Speedfan" that I have been using to monitor GPU/CPU temps, but which might have been conflicting for gpu fan control (nope.)
Having integrated graphics on or off in motherboard settings.
I should mention that the 3070 is now the only component that isn't new, from a previous build. I think it's the card literally dropping out.
Update 2 (Fixed):
Faulty graphics card, fairly certain, borrowed a 3080 from work and I've not had the error yet today.
This was a common issue 10 years ago, and still is to this day. RAM is a bulk production item because we all need 1 2 and 4 dimm depending on our budget. Next time you have a bunch of issues, look at the manufacturers RMA rules. Most of the time it's 12 to 18 months from time of purchase. Plenty of time to burn out faults. Never purchase what you don't have to, RMA when you can.
Update 2 (Fixed):
Faulty graphics card, fairly certain, borrowed a 3080 from work and I've not had the error yet today.
Hey there, is the system with the new GPU still going stable?