I have a RTX 3090 FTW3 ULTRA GAMING. When I started my computer today, I was greeted with some small boxes of various colors. Some came and went, but didn't come back. Some flickered. Some flickered off as I hovered the mouse over them. They have since went completely away, but now I am concerned that the video card is going bad. Is there a way to test the VRAM?

Also, my GPU was cold. I had just booted it up. There was no overheating going on.

glenn37216
Could be possible bad card... But still also could be other things like psu not giving gpu enough power on 12v rails. The test should show some type of anomaly if something is failing. . ..( gpu drivers can also cause this corruption)

There was no load on the GPU, so I doubt it was the PSU. Especially, because after this I booted up a game and it ran fine hitting close to 99% GPU usage.

You can use the program occt to test the vram.

was it new or used

Sajin
You can use the program occt to test the vram.

How long do you run the test for?

Sajin

chrisdglong

Sajin
You can use the program occt to test the vram.

Thank you Sajin. Is this the site where you get OCCT from: https://www.ocbase.com/
 
I haven't used this tool yet.

Yep. An hour runtime should be good enough.

I haven't had the time to run it for an hour yet, but I did run it for 30 minutes and there were no errors. I am wondering if any software could have caused this? Icue crashed around the same time... I know this one is probably a long shot. My other thought was the monitor, as it uses 2 cables to get 4K 144Hz. Maybe it was a glitch with that not so normal solution? If I remember correctly, the small boxes were only on one side of the monitor. This monitor patches two connections together, each displays a side of the screen. I have the XB273K from Acer.

Must have been a glitch.

...or a possible glitch with the monitor cable. Make sure you are using a high quality cable. 

Thanks everyone. I will probably just keep using it and see if it happens again. It really did give me the feel of a software glitch more than anything. It didn't seem like the monitor because I could hold down the mouse and pull the selection highlight across the area and the issue would disappear. Usually, a monitor glitch (in my experience) would not respond like this. Windows has had issues with the dual cable setup in the past. Resident Evil Village hated running in full screen mode on this monitor, it glitched out when I tried selecting it. Windows will not properly wake up from having the monitor asleep with dual cables. I essentially get the same issues that anyone using a multiple monitor setup would. But, that 4K 144Hz 10 bit color sure is worth it.
 
That OCCT tool is actually really great. I'm not sure why I haven't used it before this situation. If nothing else I got a great new tool to use in the future.

It came back today. I turned off the dual cable setup and it went away. It seems that it is either the Nvidia drivers (I did update these recently) having issues with the dual cable display, maybe a cable, maybe the ports on either the monitor or video card, or Windows having an issue with the setup...
 
 
My computer crashed while running OCCT today for the GPU, not memory, picture of the memory dump is attached.
 
Edit: I think this is a driver issue. Every time that I restart my computer the issue goes away. When the issue was happening last, I ran OCCT to test the GPU and it crashed with the error in the image. After the crash, and my computer restarted, I ran OCCT for an hour straight without issue. When I restart the strange squares are gone as well. When i tried to simply turn off and back on the monitor, the issue was still present (ruling out the monitor).
 
I notice a lot of people had BSOD and crashing issues with 466.47, so I will DDU in safe mode and install the latest drivers and see if the issue goes away.
 
EDIT 2: I think this whole issue might have been related to EVGA Precision. I installed new drivers. I started a game and the GPU clock was stuck at 210MHz. I tried multiple games and the same thing was happening. I uninstalled Precision and everything works fine. I can't for the life of me understand why EVGA threw out Unwinder for the people they have working on Precision now. Every time that I have tried Precision after Unwinder, it has always resulted in at least disappointment, at worst pure frustration.
 
My thinking about the original issue is that the GPU clocks were stuck at 210MHz randomly because of Precision. I'll leave Precision uninstalled and see if the problem resurfaces.
 

Sajin
Starting to sound like a gpu issue now.

I think if it happens again, then I will have to contact EVGA for an RMA. I'm hoping either uninstalling Precision or the fact that I installed new drivers (466.77) will fix it.

Issue came back. It took a while because I haven't been using this computer much lately. I am wondering if it has to do with the dual cable monitor setup, so I set the monitor to only run on single cable mode (120Hz at 4K with no 10 bit color instead of 144Hz at 4K with 10 bit color). I will see how this goes. This issue never shows up in games, only every once in a while right after booting. It will not go away until I reboot.
 
EDIT: One of the two cables might be bad as well. I will rule out bad cables as well.
 
EDIT 2: I believe this issue was software related. I say this because it never does this anymore. It just went away. I'm sure that it was either Windows or Nvidia drivers causing the issue.

So the issue never went away in single cable mode?