Freezes then restarts nvlddmkm.sys

jds63 · Jan 17, 2023

Ran 3D Mark these are results for graphics stress test

StarDude · Jan 17, 2023

Here's something interesting. I had a few crashes this morning playing RDR2 which is the first time since I posted my fix. I did something differently today than I did the previous 2 days. I used ALT-TAB to switch into Windows Edge to check some web links while paused in the game and then ALT-TAB to go back to the game. Doing this appears to add instability to the nvlddmkm driver. My TdrDelay is set to 2 seconds so I'm going to change it back to 10 seconds and see if I have any issues with ALT-TAB.

So if this is a trigger for nvlddmkm, that would explain why all the benchmarking tests work out perfectly for everyone. I'm also going to try running MSI Afterburner and do some other benchmark tests and try ALT-TAB in the middle of them.

squidbats · Jan 18, 2023

Hey there, I joined to say I have the exact same problem, been a few weeks of following different threads everywhere, and it seems like most of the current fixes that have worked for some have made their way here.

I have black screen crashes every few hours in games, usually DX12, on 528.02, although possibly slightly less frequently on the 528.02 Studio drivers. Im probably going to go back to 516.93 as I got the crashes once every few weeks on those drivers, strangely they would be a frozen screen rather than black screen, but would still pop up in event viewer as nvlddmk errors.

So far I have tried:
Enabling Resizeable-bar
DDU new drivers
Disabling XMP
Putting computer on seperate power socket
underclocking 200mhz on the GPU
Swapping out ram
Swapping out PSU
Making sure my PCIE settings arent set to Auto
underclocking RAM
updating BIOS

Studio Driver and ReBar maybe? Decreased the frequency of crashes slightly, but its probably placebo.

Kind of at a loss now, plenty of people with the same problem tried RMAing their cards and had the errors continue so its unlikely its hardware for all of us. Im going to try the registry tweak now, and then probably go back to the 516.93 drivers from a few months ago as every newer driver has made it worse.

Im on W10 btw, Ryzen 5 3600, 3060ti.

Cromax · Jan 18, 2023

squidbats said:
Hey there, I joined to say I have the exact same problem, been a few weeks of following different threads everywhere, and it seems like most of the current fixes that have worked for some have made their way here.

I have black screen crashes every few hours in games, usually DX12, on 528.02, although possibly slightly less frequently on the 528.02 Studio drivers. Im probably going to go back to 516.93 as I got the crashes once every few weeks on those drivers, strangely they would be a frozen screen rather than black screen, but would still pop up in event viewer as nvlddmk errors.

So far I have tried:
Enabling Resizeable-bar
DDU new drivers
Disabling XMP
Putting computer on seperate power socket
underclocking 200mhz on the GPU
Swapping out ram
Swapping out PSU
Making sure my PCIE settings arent set to Auto
underclocking RAM

Studio Driver and ReBar maybe? Decreased the frequency of crashes slightly, but its probably placebo.

Kind of at a loss now, plenty of people with the same problem tried RMAing their cards and had the errors continue so its unlikely its hardware for all of us. Im going to try the registry tweak now, and then probably go back to the 516.93 drivers from a few months ago as every newer driver has made it worse.

Im on W10 btw, Ryzen 5 3600, 3060ti.

Did you try maximum performance in control panel? and disabling Pcie power throttling in bios

squidbats · Jan 18, 2023

Yeah, I also get the same voltage spike just before the crash as well.

my minidump also says its a DPC_WATCHDOG_VIOLATION at IRQL of DISPATCH_LEVEL or above, which is common in other types of driver errors.

Quite a few people had identical problems caused by HW monitoring apps and RGB software such as iCUE and Armory Crate. For them removing them or at least making sure they didnt control RGB on the GPU fixed their issues, very odd.

leonardoinhoqui · Jan 19, 2023

Hi everyone, after updating the driver to version 528.02 and changing the energy manager mode: preference for maximum performance on the nvidia panel, I had no more problems. The last time it happened was on 01/11 at 9:34 am. If it happens again or not, I will update here to let you know.

squidbats · Jan 19, 2023

So after reading the reddit posts for the most recent driver release there are 100s of people with the same problem, a worrying number of them have found after updating to the most recent drivers rolling back no longer fixes the issue leading to speculation that the drivers have somehow corrupted the vBIOS on the card. Worrying if true as I dont really want to flash my GPU lol

jds63 · Jan 19, 2023

squidbats said:
So after reading the reddit posts for the most recent driver release there are 100s of people with the same problem, a worrying number of them have found after updating to the most recent drivers rolling back no longer fixes the issue leading to speculation that the drivers have somehow corrupted the vBIOS on the card. Worrying if true as I dont really want to flash my GPU lol

Can you post articles here ?

StarDude · Jan 19, 2023

I checked and mine was already set to Maximum Performance. Here's a screenshot of the nVidia control panel.

squidbats · Jan 19, 2023

So after rolling back to 516.93 mine is far more stable, so perhaps those people were just not DDUing properly. One really odd thing I had 3 crashes that were almost identical after downgrading, but they gave a different error in event viewer to do with ntkrnlmp.exe, which turned out to be caused by a driver for hamachi of all things, maybe this was to do with changing the timings in the registry? Uninstalling Hamchi fixed it, I also used the uninstall tool for Armory Crate to remove any extra things lurking left over from when I had it a year ago.

Either way this is 100% a driver issue on Nvidias end, perhaps to do with display port multiple screens and windows updates conflicting. Seen a lot of the people having the problem have multi screen set ups. Who knows, I guess im staying on this ancient driver for the time being.

jds63 · Jan 19, 2023

squidbats said:
So after rolling back to 516.93 mine is far more stable, so perhaps those people were just not DDUing properly. One really odd thing I had 3 crashes that were almost identical after downgrading, but they gave a different error in event viewer to do with ntkrnlmp.exe, which turned out to be caused by a driver for hamachi of all things, maybe this was to do with changing the timings in the registry? Uninstalling Hamchi fixed it, I also used the uninstall tool for Armory Crate to remove any extra things lurking left over from when I had it a year ago.

Either way this is 100% a driver issue on Nvidias end, perhaps to do with display port multiple screens and windows updates conflicting. Seen a lot of the people having the problem have multi screen set ups. Who knows, I guess im staying on this ancient driver for the time being.

Your problems are different then mine, yours occur in games and receiving different errors then me. So not sure going back to this driver would help since out way before i even had my Graphics card.
NVIDIA Studio Driver | 516.93 | Windows 10 64-bit, Windows 11 | NVIDIA

squidbats · Jan 19, 2023

jds63 said:
Your problems are different then mine, yours occur in games and receiving different errors then me. So not sure going back to this driver would help since out way before i even had my Graphics card.
NVIDIA Studio Driver | 516.93 | Windows 10 64-bit, Windows 11 | NVIDIA

On the newer drivers I have the exact same error as yours but in games, nvlddmkm.sys black screen
\Device\Video3
Error occurred on GPUID: 100.

Mine only started after I updated from 516.94 (I meant 94 not 93 as thats the game ready driver) to more recent drivers. So the next few drivers after 516 might be okay I just havent tried them.
It cant hurt to try anyway? Just make sure you use DDU.

jds63 · Jan 20, 2023

squidbats said:
On the newer drivers I have the exact same error as yours but in games, nvlddmkm.sys black screen
\Device\Video3
Error occurred on GPUID: 100.

Mine only started after I updated from 516.94 (I meant 94 not 93 as thats the game ready driver) to more recent drivers. So the next few drivers after 516 might be okay I just havent tried them.
It cant hurt to try anyway? Just make sure you use DDU.

My problem started in beginning of November so that driver then till newer driver now was ones i had used, till now was still occurring. Mine is intermittent. Right now mine has not done anything since 1/11/23.

Nuku · Jan 21, 2023

Hey all, reading this thread has been very insightful as I've been having potentially the same issue. Still haven't found a solution and I'm on WINDOWS 10(!!!) but I wanted to share some of the logs/observations/steps done because I have a feeling this isn't really down to Win10 or Win 11 and maybe it helps anyone identify the isuse. If not good, please just remove or ignore!
My rig was a pre-built one and is about 4 years old. Apologies if something doesn't make sense - I work in Tech but I'm not crazy well-versed with anything in the firmware/driver world and have superficial knowledge regarding most hardware areas

My Rig

Symptoms:

Screen freezes regardless of system load (happened when gaming, when completely idle, when only having base apps like browser + Spotify open, or also while not doing anything load intensive)
Fans, drives, motherboard etc all still running - other evidence of activity: Sometimes sound will still play for a while, or sound will loop the last few seconds for a while, or sometimes when pressing buttons it will still make Windows sounds showing some OS functionality still running just graphics being dead(?)
Cold reboot needed to recover
Sometimes display won’t come back after cold reboot - dual monitor setup only showing “no signal” while fans and other hardware are running/LEDs on/showing activity - in that case I usually have to unplug one or both of my monitors from the GPU (1x HDMI, 1x DPort) and try another cold reboot. Can plug both monitors back in afterwards
Sometimes there will be a few seconds of stuttering frames and audio before the crash
Sometimes there will be a Windows pop-up that app.exe (e.g. Overwatch, Discord, Firefox) has been blocked from accessing hardware. This may not always lead to a full crash but the app affected will crash
nvlddmkm driver errors shown in Event Viewer in some cases right before the crash or before the Windows pop-up
Other times there is no log evidence in the files, last log entries are usually a few minutes before crash
I noticed that while temperatures/power usage doesn't seem to be abnormal, logs for GPU show some weird data values moments before the crash happens (e.g. temp logged as 0 so assume the machine doesn't receive any data from sensors) and logging resumes normally after a successful reboot. I use OpenHardware Monitor. See screenshot of data in question from a crash that happened today (21/01/23).

Eventviewer logs for the same time stamps:

Troubleshooting done so far:

Completely reinstalled drivers multiple times, including the use of DDU from safe mode, multiple versions of Nvidia drivers over the last 6+ months
Updated drivers for pretty much anything else available
Opened up my computer and checked that all parts are seated correctly, dusted off parts (my setup is designed quite well, barely had any dust!)
Tried a few combinations of power plugs and power sockets in case of dodgy power setups
Changed power options to high performance in case it’s due to power issues
Changed timeout + recovery settings in registry as per How to Fix ‘Display Driver nvlddmkm Stopped Responding’ on Windows 10/11
Turned off hardware acceleration in various apps incl Windows overall, browser, discord and games that use it
Reseated GPU
Checked that GPU isn’t overclocked
Underclocking via MSI Afterburner
Temperature checks/logging didnt seem to be a temp issue although GPU is running on the warmer side when gaming (Overwatch/Cyberpunk up to 83°C), after lowering graphics settings in Overwatch runs at around 50°C
Reinstalled Overwatch as it’s the game I play the most
Disabled a few gimmicks from Geforce experience (Shadowplay, ingame overlay etc)
Memtests, multiple - came back ok
Removed lots of crap/superfluous apps from PC, incl most apps that could interfere with GPU settings and that I no longer need
Calculated PSU capacity required for the setup I have: I have a 600W BeQuiet Bronze, Estimates vary from 300W upwards, only reaches 600W if I massively overcalculate everything just in case? However this does not explain crashes when the PC is almost idle?
Scannow - sometimes came back with errors but resolved itself in between, still did a few windows repairs via DISM just in case
Looked at sensor logs for temperature anomalies/power surges etc - only abnormalities are around GPU sensors, see screenshot - temperatures not higher than usual/or definitely not high enough) - but noticed that temperature logging had stopped / a temperature of 0 was logged a few mins before crash (e.g. logged half an hour of temps around 45°C for GPU, then 5 mins before crash only 0°C, then logging resumed after a few minutes when the reboot was complete) - see screenshot above

Haven’t tried yet:

Pretty much anything BIOS related - not feel comfortable
Flash GPU - dont want to risk
Complete clean reinstall of Windows
Replacing parts like GPU, PSU or monitors (might test a new monitor, want a new one anyway..)

WXC · Jan 21, 2023

@Nuku

Welcome to ElevenForums.

You may also be interested in checking out/posting at BSOD Crashes and Debugging, on the Win10 sister site. :wink:

Best of luck.

StarDude · Jan 22, 2023

My crashing is back so have been doing more testing.

I used the Microsoft Process Monitor tool this morning to capture all the process, registry, network and file activity that are happening while a crash happens.

Process Monitor - Sysinternals

Monitor file system, Registry, process, thread and DLL activity in real-time.

learn.microsoft.com

I can see Windows starting up the Window Error Reporting process: PID: 30708, Command line: C:\WINDOWS\System32\svchost.exe -k WerSvcGroup

I went looking for these error logs in C:\ProgramData\Microsoft\Windows\WER\ReportArchive

All the crashes are Kernel 141 errors which refers to a hardware failure. That lead me to this thread which is a lot of the stuff I've already tried to no avail. Tried the PCMark10 test and it came through at the 97 percentile rating with no errors.

Redirecting

Still thinking it's a video card issue or a Windows issue dealing with a game and other stuff running in the background, but don't have the luxury of having a second expensive card to swap out and try.

jds63 · Jan 23, 2023

StarDude said:
I went looking for these error logs in C:\ProgramData\Microsoft\Windows\WER\ReportArchive

All the crashes are Kernel 141 errors which refers to a hardware failure. That lead me to this thread which is a lot of the stuff I've already tried to no avail. Tried the PCMark10 test and it came through at the 97 percentile rating with no errors.

What else shows in Event Viewer>Administrator Events at time and dates of these errors besides nvlddmkm.sys?

Don't think it's hardware failure since works okay in Windows (No black screens ?). What about other games besides one in question ?
Think it's seems to be RDR2 game related issue only.

Also no nothing of your system specs!!!!!!!!! Try listing them in your profile here

StarDude · Jan 23, 2023

So here's something interesting I tried yesterday. My motherboard is Gigabyte so I downloaded the Gigabyte Control Centre. In there is said the latest VGA driver was version 496.13 which is pretty old. I've been running the latest nVidia drivers 528.02. I figured I'd give 496.13 a try since nothing else has been working well so far. I did not use DDU to do a clean removal but installed over top of the 528.02 drivers.

What I experienced was that my VR headset would freeze for 5 to 10 seconds but never crash. My event log is clean and I played for a couple hours without a single crash. That makes me think the TdrDelay is actually being used with the older drivers and that 528.02 ignores that setting (or something to that effect).

Now that's not all of it. My 3DMark benchmarks have dropped by 25% with the older drivers eventhough the clocking speeds have not changed. Is it possible that newer functionality in the current drivers gives a good performance boost but also introduces a crash state?

Still more testing to do......

StarDude · Jan 23, 2023

The crash errors I get in the Windows Event logs are the same as everyone else.

jds63 · Jan 23, 2023

Do you have a dedicated Graphics Card ? Or Integrated one on motherboard ?

Your still not listing all system specs Hardware

Freezes then restarts nvlddmkm.sys

Well-known member

My Computer

New member

My Computer

Member

My Computer

Well-known member

My Computer

Member

My Computer

Member

My Computer

Member

My Computer

Well-known member

My Computer

New member

Attachments

My Computer

Member

My Computer

Well-known member

My Computer

Member

My Computer

Well-known member

My Computer

New member

My Computer

Well-known member

My Computer

New member

My Computer

Well-known member

My Computer

New member

My Computer

New member

My Computer

Well-known member

Do you have a dedicated Graphics Card ? Or Integrated one on motherboard ?​

My Computer

Similar threads

Do you have a dedicated Graphics Card ? Or Integrated one on motherboard ?