ROG-STRIX-RTX3080-O10G-GAMING: nvlddmkm.sys BSOD

[PipJones]

Hi,

GPU: ROG-STRIX-RTX3080-O10G-GAMING. This is in an 10900X / Strix X299-E II system.

I have been having issues with a random Windows 11 BSOD naming "nvlddmkm.sys" as the culprit.
- occurs when using windows 11, no pattern
- occurs in gaming, only tested with WZ so far.

I have not been able to spot a pattern yet. Fresh install of W11, latest nVidia drivers etc. No software overclock application installed for GPU.

The problem appears to go away when I run the card in "normal" mode (i.e. the physical overclock switch on the card is "Off").

Is anyone else out there getting BSODs with this card?

Does anyone have any thoughts on why the physical OC would induce a BSOD?

[Ribellu]

I have this error when I play in 4K by setting my screen to native 144Hz.

And if I remember correctly this error corresponds to the driver who is stuck in an infinite loop a few seconds before crashing, because he is no longer fast enough to respond to the data.

By looking on the internet for two options each came back to either the OC too strong, or playing at a very high resolution with a very high screen refresh in my case 144Hz and 3840X2160, I blocked it at 120Hz and I didn't never had this error again.

PS: need a very powerful GPU to be able to play any game in 4K and with a very high screen refresh rate, some will work because some graphics engine uses more the CPU than the GPU, but in general it is rare, in case of doubt with this error it is better to lower the refresh rate of the screen to just below its maximum to play in 4K.

[PipJones]

Agreed.

Any chance you could direct an @Asus person to this thread and ask them to verify what I have been told in chat?

It would be nice to know what it would cost to have the card professionally analysed and what you get for your money etc.

[Nate152]

If that is what ASUS chat support told you, I would hold it to be true.

I can direct you here, there is a phone number for the U.K.

https://servicecenter-nearme.com/asus-support-uk-customer-service-number/

[PipJones]

Thanks Nate.

If and when it gets to the point where i think i have a hardware fault on the card i will go through that pain barrier

I'm still quite a way off that. I need to re-do a W10 install first.

[PipJones]

Small progress report. RL got in the way.


I disabled 2x peripherals in BIOS: WiFi and 2nd onboard NIC and W11 latest updates applied.

After about 5 hours testing in WZ, No BSOD - but I did get what I believe to be the "usual" WZ crashes.

Error code: 0x0000000143cf3197 10828030 0xc0000005
ModernWarfare.exe





I can't draw any conclusion from that other than "no BSOD is encouraging".

Fresh W10 build is also ready to go.

[Nate152]

Nice, the driver not crashing is getting much better.

Here are a few fixes for fatal errors in WZ.

[IIINikolaiIII]

Hi. I'm facing these same issues you have been dealing with, but with a 3070TI Strix OC. FH5 and Halo Infinite are crashing constantly either with a BSOD or with a driver crash. Event viewer says the driver crashed and recovered. This happens with stock settings. Using Win11 aswell. Never had any issues and don't have with other games. Anything i can do?

[PipJones]

Hi,

Was your system stable prior to W11?

My starting point would be to go into your BIOS and disable everything that you're not using.

These are all the things that I did on my system:

1) Fresh W11 install and all updates applied. Minimal monitoring software installed (only AIDA and SIV)
2) Update GPU BIOS
3) Update GPU driver. Use DDU and do a clean install. Set TDR delay as per earlier post.
4) Disable GPU OC
5) Disable RAM OC
6) Disable 2nd LAN port
7) Disable WiFi

I've put about 5 hours into FH5 this morning, the Goliath event - zero issues.

It's very difficult to say that "it's fixed", but, I haven't had a BSOD or CTD (crash to desktop) since #6 and #7. My guess would be that it's a W11 "thing". Something changed under the hood in the most recent updates.

[PipJones]

@Ribellu: All points noted r.e. 120Hz. I did test. Nothing to report other than it didn't crash, but, it was a very short test.

I still have issues and do do not have an answer. I am posting to document the journey and what I have learnt along the way.

1) With the Factory OC on: Hang in FH5 within 30 minutes or so. No BSOD or CTD, just a game hang.
2) With the Factory OC on: W11 BSOD, no gaming. Just opening XLS to look at logged data.
3) With the Factory OC off: Eventually got the same "hang" in FH5


AIDA logging is showing that flicking the factory OC switch is not doing what I expect. My card reports maximum GPU clock is 2040MHz with the switch in either position.

I am now testing a forced downclock using nVidiaInspector.

[Ribellu]

Have you tried a "clean" uninstall of your driver with DDU in safe mode, and then reinstalled?
Sorry if you have already mentioned this before but I have not read all the pages in this thread.

I block my refresh rate in the driver: