Freezes then restarts nvlddmkm.sys

Hello, I'm glad I found this thread. First of all, I'm from Germany and my English isn't the best.
Yes, I also have the problem that my system freezes, but only when it is idle or with browser applications.
When I load my gpu e.g. I can play Farcry 6 in 4K and maximum settings for hours without any problems. No problem for the GPU.
I have 2 monitors, one connected via DP and one via HDMI. I entered the rest of my hardware in my profile

What I've tried so far:

-Windows new install
-Windows SSD renewed (before Samsung 690 EVO M.2 SSD 512GB @ PCIex3 Now Samsung 980 Pro M.2 1TB @PCIex3)
-GPU Driver renewed (the nvidia Driver)
-PSU renewed (before beQuiet dark power pro 850W Now Corsair HX1000i 1000W)
-Installation of the old GPU GTX1080, the error never occurred, so I know the error is in the 3090
-XMP on or off
-different Core Clock Ratios (Stock 4,3GHz)
-Fastboot is disabled

What I haven't tried yet:
-Flash the GPU BIOS the latest is installed
-Firmware of the GPU renewed the latest is installed

Other software that otherwise runs in the background:

-HWinfo64 pro
-Elgato Streamdeck
-iCue
-Zotac Firestorm for the GPU settings
-Logitech GHub

I also tried the other tips like changing the energy mode in windows and Geforce.

Oh and i have Win 10 installed yet.
I hope this is okay.

Have you tried the TdrDelay registery update? I put this in place a week ago and haven't had a crash since with my 4090:

1. Goto registry editor (start -> run -> regedit)
2. Navigate yourself to HKEY_LOCAL_MACHINE/System/CurrentControlSet/Control/GraphicsDrivers (make sure GraphicsDrivers is highlighted).
3. Right click -> New -> DWORD (32-bit) value -> Name it: TdrDelay -> double click it -> Set it to A (hexadecimal) or 10 (decimal).
4. Reboot your PC

FornaxLacerta said:

Have you tried the TdrDelay registery update? I put this in place a week ago and haven't had a crash since with my 4090:

1. Goto registry editor (start -> run -> regedit)
2. Navigate yourself to HKEY_LOCAL_MACHINE/System/CurrentControlSet/Control/GraphicsDrivers (make sure GraphicsDrivers is highlighted).
3. Right click -> New -> DWORD (32-bit) value -> Name it: TdrDelay -> double click it -> Set it to A (hexadecimal) or 10 (decimal).
4. Reboot your PC

Click to expand...

yes is set when i found this Forum 3 days ago and today i had 2 freezes again

Hi! glad I found this NEW thread since all of the other were kinda old.

I have had the same issue for the past month and there is just no luck. I have tried everything so far.

Some people pointed out is a RAM issue even tho memtest86 has no errors. I dont have more rams to test that but I post this in case it helps others.

I will try new suggestions that I saw here and will be posting back.

Good luck to everyone with this annoying issue, if somone find fixes pls always post your fixes and don't leave us hanging in this awful situation.

My specs in case it helps: (It is a prebuilt)
i9 12900k
RTX 3080 Aorus OC 10GB
2tb nvme
64gb ram 5200
XMP on.

BTW quick edit. I tweeted to nvidia (Im desperate) they probably wont pay attention, but if you guys comment or something on this threath maybe they will )

Hi guys!
Following this thread for a while now. Registered to share my experience and solution I discovered.

TLDR: Underclock GPU, and undervolt the card.

Lowered the Core Clock by 100MHz (~10% speed of my card) and set Power Limit of the card to 98%. Used MSI Afterburner 4.6.4.
At the moment of writing it is almost 2 weeks without crashes. During this period I haven't noticed any performance drops or issues.


Lowering the clocks. No idea how much you need to lower clocks on other cards. I'm assuming the amount would be different on different cards. My guess, slightly lower than peak performance.

Lowering the power. Same here. My guess, setting power limit just slightly lower than peak performance would do the trick.

Drivers. Tested 522.25; 528.02 and 528.24. All of them crushed. On 528.24 it crushed daily and sometimes several times a day. After applying the solution I had time to test only 522.25 (the older one), which works absolutely fine now. If old driver previously crushed and now it is working, I think it is reasonable to assume it might work with new drivers as well. Definitely going to test new drivers soon.

Running old system here:
- i7 4790; Z97; 16GB DDR3; GTX970 4Gb; 750W; Win10Pro 2604.


PS. I won't write the long list of things I tried before I got to this solution. I assume I've tried everything I could find online. Needless to say non of it worked.

I'm seasoned PC user, but not an expert in any way, and have no idea what is really happening and why. But the result is clear as day. Card is able to perform tasks it failed to perform two weeks ago. And I couldn't be more happier about it.


I hope my experience will help other on their path.
Good speed and thanks for the things I learned from all the posters on this thread.

Sadly, this issue for me was completely 100% the GPU, I thankfully had bought at microcenter and had the replacement warranty.
No longer am I crashing randomly.. I paid a little more and have a 4070TI and the system is 100% stable now.

Prior, I was getting crashes sitting at desktop, quickbooks, etc.
Tried DDU, and many other "fixes"

Mine was a bad card.
I had an Asus Tuf 3080 12gb model.

Hello everyone, looking to hopefully add to this very helpful thread with a possible solution.

Been having the same (or very similar) issue as OP for the past couple months with a brand new build and have been following this thread. System only crashes when doing basically nothing or after waking it up from sleep/hibernate and never under heavy load or gaming. My crashes would seem to happen almost once a day if not every other and I would have the same nvlddmkm.sys in my event logs.

Does anyone else that has this issue always have HWinFO64, Fan Control, or any other hardware monitoring app always open and running? Since the crashing seems to be related to a TDR issue with the GPU it made me think what is constantly making requests to the GPU. So I stopped running HWinFO64 and Fan Control, and I haven't had a crash since. Not sure if this is purely a coincidence or not so I'm curious if anyone else is also running these applications as well. I haven't made any changes to my system, not even increasing the TDR time as many have suggested.

System Specs:
Windows 11
AMD Ryzen7 5800x3D
ASUS ROG STRIX B550-I GAMING (latest BIOS)
Nvidia 4080FE (driver 528.49)
DDR4 3600MHz 2x16Gb
32" 4K144hz, 27" 4K60hz

Appreciate everyone's input on this thread, its been a huge help.

Sigh, another crash, gonna try to undervolt

Another update: I have had a full week of stability with my FE 3080. This is what I did: checked the latest bios from MSI for the motherboard because they had issued a couple of similar though different ones but have now consolidated them to the latest one which I installed; put my 3080 back to stock settings (from under-volting it); set the cpu to high performance mode in Windows 11 power settings (it still drops down the voltage when less power is required); set 3D in Nvidia Control Panel to be balanced and ran DISM /Online /Cleanup-Image /CheckHealth (which found 4 component services corrupt and repaired them, not found with sfc /scannow) and installed NVIDIA Resizable BAR Firmware Update Tool. I have the latest studio driver and version of Chrome. At this point I cannot say which change/s fixed it but if I have time will reverse some of the changes to see if it becomes unstable again (it was only ever unstable using Chrome when browsing quickly and occasionally using YouTube and was reported mainly as Windows was not shut down properly. The motherboard is an X299 (PCI Gen 3) and the graphics card is on the 2nd x16 slot.

I started this thread back in November since then my issue has been intermittent till this day.

Have not had an issue in almost a month and a half now and on 2/25/23 it happened again.
To explain what i was doing when it occurred, not that much ! I was on the internet using Firefox and was doing a lot of left clicking on my trackball when it happened. It froze and i could not use my trackball or my keyboard to get out of it.
PC did not restart on it's own i waited, so then i pushed restart button on my case.

This is what i received in Event Viewer as usual like before in past.
I appreciate all the similar issues posted here and know i am not alone on this as i was in the beginning. Some of your issues posted are exactly same, some of more consistent of an issue occurring more often or within gaming, etc...

Kernel 41 (63) i get at the end of my Event Viewer is a cause by me having to restart PC because of no response from keyboard or trackball to get me out of it. Meaning it was not cleanly shutdown so that critical error happens.

nvlddmkm.sys error is common error to be shown in cases when a video error has occurred. Many reasons as we know this error can be shown.
Also meaning that just because we all have it in our Event Viewer doesn't always mean our problems are exactly the same.

I am always on the newest NVIDIA drivers as they come out to let you all know. Will keep this topic open till we all find a cure.
Somehow feel it maybe Windows 11 related for most.(Although very few posted using Windows 10)

Hey All, see specs here:
Windows 10 Pro
I9-13900k
MSI Pro Z790-A Wifi Mobo
32gb Corsair Dom Plat RGB 5800mhz
1000w EVGA Super Nova 80+ Gold
Zotac Trinity OC 4090
Samsung 980 M.2 2TB (OS and Storage)

-PC was stable for two weeks after building.
-Randomly started seeing crashed with the nvlddmkm.dll errors reported here in Event Viewer.
-Crashing mainly during games (Forza 4, BF 2042, Undisputed, pretty much every game) and once at desktop
-Mouse or game freezes completely and usually PC restarts after a few minutes if I dont restart it myself.
-All parts were purchased from MicroCenter.
-All drivers and Windows updates are in place.
-Power management settings are all set to Max Perf. Did not see any improvement.

Beyond unhappy at this point. $1,700 GPU and this is what we end up with....
Have been an NVIDIA Geforce fanboy for 15+ years.... this may change that if a solution isn't presented by Nvidia soon.
Very painful to see so many accounts of this with no response from the manufacturer anywhere.
The issue has been present for a LONG time for many users and has persisted through multiple driver updates.

Curious if anyone has RMA'd....and if they have switched to a different card with success?

I am still trying to determine if its a hardware issue or a driver issue. At this point, I am almost convinced its faulty hardware.

RMAing cards has fixed it for some but in my experience of researching it by far the least. Mostly seems to be RAM or RAM slots failing on the hardware side of things, for me it was my entire motherboard. Replacing it with a board with better VRMs has completely fixed my problem, but it took my previous board literally burning out during a stress test with a known good GPU to find that out.

Hey there, turns out, final desperation attempt in ripping out one of my DDR5 dimms, ( was running dual channel slot 2 and slot 4) - this fixed the issue completely. It got so bad that I could not play a game for more than 10 minutes with crashing (with artifacting). Moving away from dual channel weirdly enough stopped all crashes. My specs are posted in my last post if your interested. Check your DIMMS people. These newer GPUs are so finnicky, then throw in a Z790 chipset, DDR5 Ram, and ridiculous power/thermal requirements and so-so software support…. Recipe for issues. I have been building and troubleshooting PCs my entire 31 years in existence, and its almost ALWAYS RAM or PSU causing issues despite “obvious” symptoms. Its super rare that just a GPU is the root cause. Also, set your BIOS to default settings, adjust fans and leave it alone. Those of you with OC and tuned voltages that are wondering why you are having problems… well there is zero need to overclock this new hardware. If you do, you will have issues. Keep it stock and the complexities of troubleshooting reduce dramatically.

I can post in more detail after 5pm EST tonight if anyone has any questions. Sorry for being vague, multitasking at the moment.

Ram Configuration during crashes:
32 GB - 16gb x 2 - Corsair Dominator Platinum RGB - 5800mhz - DDR5 - Slots A2/B2

Ram Configuration with crashes resolved:
32 GB - 16gb x 2 - Corsair Dominator Platinum RGB - 5800mhz - DDR5 - Slots A1/A2

Pulling a DIMM does not mean for certain that the RAM was the issue. What it does mean, is that it is not a faulty GPU. Most likely my sub-par main board having some firmware based conflict between RAM/VRAM/GPU. Could also have been the BIOS settings associated with having the high speed RAM in place, and it is possible that pulling out one of the DIMMs reset whatever BIOS setting was causing the crashes. Either way, an RMA was not needed, and was narrowly avoided along with a giant headache. Check you RAM config. Check your BIOS config. Set everything to default config and work your way down from there. That is how you rule out software issues. If things are breaking with default configuration, it asserts a hardware issue.

I knew this build was going to be rough because its all BRAND spanking new tech/hardware. I was mentally prepared to have expensive stuff not work as expected. Luckily the fix is what it always is…. RAM config/compatibility issues hiding in plain sight.

Additionally, Looking in your operating system logs (Windows Event Viewer) for these types of problems is going to lead you into a ton of red herrings and false flags. Most of the time, persistent crashing (Especially ones that involve any type of visible artifacting at the time of the crash) is not caused by an application layer/operating system layer issue. This is typically always going to the hardware itself (RAM, MOBO, GPU, PSU), the configuration of said hardware, or BIOS settings going awry (or a combination of all of them).

TLDR:
Issue was narrowed down (and resolved) to the below possible issues for others still troubleshooting:
-Bad RAM config (Make sure your clock speeds are what they are supposed to be)
-Bad Motherboard (HATES dual channel) (No crashes with memory running in slots A1/A2, as opposed to A2/B2)
-Bad BIOS settings (Set everything to absolute default, when in doubt, AUTO is the setting to go with)
-Incompatibility somewhere between RAM (In dual channel), GPU, and BIOS settings.

Bottom Line: Taking RAM out of dual channel, and setting all BIOS settings to absolute default config got rid of crashes.
If all else fails, replace your motherboard....not your GPU. I sure will be. Having to run RAM in a non-optimized config as a work around is annoying but better than having to RMA a $1,700 GPU.

I really hope to help people having these issues. It was very painful for me to deal with as I had spent so much money and time on the build, and seeing how widespread the issue was with no concrete resolutions was very disheartening and if I can help others in not having to deal with that stress, I will do so.

Okay, here Is an update:

I have had 3 days without a crash now. What I did was Underclock my GPU core and VRAM. But I made SURE it was underclocking. For some reason when I tried to underclock with software like MSI Afterburner or AORUS ENGINE, the clock actually didnt change. I was monitoring this with HWinFO. If I moved the slide let say -15, -25, -50 or even -100, the clock will stay the same. It was until I moved like -272 MHz that the clock actually went down to 1800MHz in (even tho msi or aorus engine told me it was at 1633 MHZ????). As for the vram clock I went down all the way to "-754" Even tho the REAL change was just like 100 Mhz at most. For now I havnt had any crashes.

I also added most of the advices here. Like Nvida power in performance mode, turned on "optimization for windowed games". Checked RAM to be OK with Memetest 96, removing hardware aceleration from chrome and discord. Disabled XMP (note to this, I woulnt have disabled XMP but for some reason it kept my pc from going to sleep. I disabled because of that and since my RAM kept running at 5200 mhz, but I think this is another porblem, not related), etc. TBH at this point I think it was the REAL underclock that helped me. Cuz It was until I did it that actually made a change.

EDIT: Also, if you have MEmetest Errors at test 6 or 7. Changing the voltage of the RAM to be 1.36 fixed all of those issues.

I will keep posting for updates. I also just downloaded (today) the latest nvidia drivers and no crash. I hope my clock fixed the issue.

Good luck everyone, I know this is a nightmare.

Another update: Since my last changes I have had 15 days of no freezes/automatic restarts when using all browsers but primarily Chrome. It also feels smoother now and I can scroll as fast as possible, jump from page to page etc. Games have never been an issue. Since many people have been referencing RAM as a potential cause I have increased the speed and tightened the latency (4 slots of 8Gb) to stress it out a little but it is still stable (now at DDR4 1.36v @ 3200Hz and CL 14.14.14.34). Will give it a further week and then reverse a couple of the changes (am now only left with lowering performance settings in Nvidia and Windows) to see if it reproduces the problem. All this happened when I replaced an MSI 1080ti with an FE 3080 at a time when Windows 11 had the 22H2 upgrade, so do not know if the latter had an impact. The only other change was to position the 3080 in the 3rd (PCI x16) slot instead of the second (x 8) slot. I have used the different PCI slots before without any problems with the 1080ti so am ruling that change out as a causation other wise I would put the 3080 in the first slot as the recommended position for a single graphics card (don't use that position as it puts the card close up to the CPU cooler and my case (Silverstone FT02) provides an option for better cooling for the graphics card in other slots due to the upward 3 fan cooling). I even wondered if the vertical position of the FE 3080 was unfavourable due the design of the FE card cooling but temps are normal and idles around 22C with set 30% fan speed). My PSU is a Seasonic Titanium 850W.

Update: I had like 2 weeks without crashes, but today it crashed while using Unreal Engine (And a bunch of other 3D programs). So even tho it was under stress, it wasnt as much as for it to crash. I will keep monitoring, if the issue persist I will underclock even more.

Update: I have had a full month without any problems and my PC is functioning flawlessly. I reversed performance changes with no recreation of the freezing / automatic restart when using Chrome. Every setting is set to stock (though have left the RAM having a higher speed and tighter timings) with hardware acceleration turn on. My problem was not in any way hardware related. The only changes that I made that could have rectified my PC's problem was the updating of the Bios (I was using the penultimate firmware) and the running of DISM /Online /Cleanup-Image /RestoreHealth which found a couple of component stores to be corrupt (SFC /Scannow found no issues). Am pleased that the issue is rectified but not so pleased that I cannot categorically say what caused it and by default not really help anyone else.

I got the same error, nvlddmkm.sys black screen
\Device\Video3
Error occurred on GPUID: 100.

I am using RTX 3070Ti, but this happened when I started Aid64 while I was playing.
But I disabled HAGS and I'll test again for a few days.

Same problem as everyone, Tried everything mentioned here.

\Device\Video3
Error occurred on GPUID: 100.

Tried everything mentioned here.

Seems there could be tooo many things causing this. I am not nearly good with computers as most of you, so will w8 a bit hopefully someone finds a fix. Also as a last resort will try to underclock GPU, but bit afraid to do it since I have near to nothing of knowledge in this.