Laggy boots, Faulty GPU Driver, Random Restarts

[Reeve2921]

Hi,

Been having an ongoing battle with a new build i did in may and hoping to get some insight to some fixes or what the issues is. Ive currently been in contact with sending the card off for RMA.

Ryzen 7 7800x3d w/ Lianli Gallahad II SL-INF LCD
MAG x670e Tomhawk Wifi
4080 Super Zotac AMP Extreme Airo
2 X 32gb DDR5 Corsair Vengeance @ 5600MHZ (was higher but dropped when issues started to occur)
Samsung 990 m.2 2TB
Corsair HX1000i Platinum
Windows 11 64 bit


The PC becomes rather unstable after installing any recent NVIDIA Driver. The PC will not run any game in full screen mode, crashes, freezes, BSOD and sometimes all screens will go black followed by a self reboot at complete random times. Its somehow broken windows on two occasions where ive ended up having to fresh install. I have tried installing only the driver from the nvidia site, Also via nvidias app/geforce. Ive also tried DDU in safe mode with no success. Eventually the PC just starts bugging out. If i try to run games in Fullscreen mode they run at 1-3 fps regardless of settings or resolution. Borderless windowed seems fine.

I keep getting Event log errors under:
NVIDIA OpenGL Drivers "Out of memory".
Another is DrvSetContext failed functionality indeterminant(pid=15636 cncmd.exe 64bit) <- this PID number changes with each error.
I get the two errors above regardless if im gaming or just sitting on the desktop doing nothing.

& The backing-file for the real-time session "NVIDIA-NVTOPPS-FILTER" has reached its maximum size. As a result, new events will not be logged to this session until space becomes available. This error is often caused by starting a trace session in real-time mode without having any real-time consumers.
Other events appear such as Kernal-eventTracing Error setting traits on Provider {8444a4fb-d8d3-4f38-84f8-89960a1ef12f}. Error: 0xC0000001 & Session "PerfDiag Logger" failed to start with the following error: 0xC0000035


Occasionally i get random self restarts. All three of my monitors will go black for a short time and then the system reboots on its own.
Also on occasion when windows boots it will instantly become extremely laggy and unresponsive right from the user login screen. I noticed in task manager i can see my gpu sitting at 100% usage from desktop windows manager & system processes which eventually freezes and forces a self reboot. I read that can be due to faulty nvidia drivers ?



Im currently using the default driver that was installed with the card as it seems to be the only one that keeps the pc stable enough to use but the same issues do persist without updating the driver just not as severe or nowhere near as frequent.

• Driver date 31.0.15.5131
• Version 31.0.15.5131

PC is fine under loads/stress tests with no issues after hours of testing. Memtests have never come back with any errors. Ive tried another PCIe Slot with the same issues/event logs so i feel its not a motherboard issue ??? Unplugged ram & all cables to reseat them. Not sure what else do. Im not using a riser cable for the GPU but i am using the supplied 12VHPRW cable from corsair which provides 600W that came with the HX1000i PSU. Temps are perfectly fine with GPU idling 30c and CPU 35-40. When gaming i dont see anything over 65c. Thinking of going back to windows 10 and seeing if stability is any different as ive not given windows 10 a chance on this new system.


Any other information wanted just ask.

 

[Reeve2921]

I have the same processor, motherboard and quite possibly the same RAM.

I also have a 990 Pro but it's a secondary drive at the minute. So the differences are GPU (7900 XT), OS (W10) and boot drive (mine is a 970).

I am not experiencing any of these issues.


So it certainly sounds like GPU driver issue from the windows event viewer log and that you are finding stavility correlates with a given driver revision.

Even though the system is a lot more stable with the out of box driver for the 4080 it still does have them just less severe.

You're also using windows 10 which makes me think to attempt yet another fresh install and just stick with windows 10 instead of 11.

I did have other concerns it might be the PSU but the fact its fine underload doing stress tests but randomly self reboots when on the desktop might say otherwise.

I may also just buy a cheap graphics card off ebay for testing purposes. Driving me mad honestly

 

[Reeve2921]

You could always use the on board graphics and see how that goes before forking out for more hardware?

Probably should have included that in the original post. I used the iGPU for about a week and didn't come across any problems. Ive just set up an RMA for the card so praying they can replicate something.

 

[Reeve2921]

Try running each stick of ram on its own.

Is there a gpu you can borrow off someone to test.

Unfortunately not, I was to quick in selling my 1080ti i had so would have to just buy something cheap to test with.

I can try running each individual ram stick on its own when the gpu is off for RMA start of the week. Im just hoping that while the card is away for RMA i dont run into anything.