Question Multicolor dots on screen

[Kurtz404]

Hello everyone !
Hoping someone can help me concerning gpu issue, NVIDIA GeForce RTX 2080

I've been experiencing lately weird looking multicolor dots, appearing most of the time in the same spot on my video games, movies when it's very dark/black.. When they show up on movies I go full screen into windowed until they're gone. It will never occur during a movie or game session if everything was fine at the beginning. I bought my gpu 3 and half years ago so i'd be surprised that it's dying.

I tried to uninstall graphics driver (through ddu) and reinstall it didn't work, my GPU temp are around 45°c while running a game.
Here is a screenshot on 4K movie. They tend to appear more on my 4K TV than on my 2 desktop screens so it's not my hdmi cables.
Thanks in advance.

https://imgur.com/a/XuAryEH

View: https://imgur.com/a/XuAryEH

 

[Newtonius]

Yeah, there's no multicolor dots visible to us in the screenshot. It's an issue with your monitor's LCD panel.

 

[Kurtz404]

If you zoom in you can see it

Yeah, there's no multicolor dots visible to us in the screenshot. It's an issue with your monitor's LCD panel.

No it's not, you have to zoom a lot on top right screenshot, above "a galaxy". Let me show you with another one.

https://imgur.com/Vn9ojLY

View: https://i.imgur.com/Vn9ojLY.png

 

[Newtonius]

If you zoom in you can see it


No it's not, you have to zoom a lot on top right screenshot, above "a galaxy". Let me show you with another one.

Ah, I see it now. That's pretty odd, especially when temps are normal and that it's happening in the same spots. The first thing that comes to my mind is that it's an issue with a bad video port on either your GPU TV. This is assuming that you've confirmed it's not your video cables or monitors itself.
And also since you already reinstalled the graphics drivers, have you also tried updating your VBIOS? (You will need an nvflash tool and a ROM file - you can usually get that from your card's manufacturer support forum/service)
Though I would save that option for last. Before doing that, I would try to check-off every possible situation - like: checking the video ports on both gpu and monitors, updating all software, checking if the GPU itself is fully functional (die and VRAM), etc.

 

[Kurtz404]

Ah, I see it now. That's pretty odd, especially when temps are normal and that it's happening in the same spots. The first thing that comes to my mind is that it's an issue with a bad video port on either your GPU TV. This is assuming that you've confirmed it's not your video cables or monitors itself.
And also since you already reinstalled the graphics drivers, have you also tried updating your VBIOS? (You will need an nvflash tool and a ROM file - you can usually get that from your card's manufacturer support forum/service)
Though I would save that option for last. Before doing that, I would try to check-off every possible situation - like: checking the video ports on both gpu and monitors, updating all software, checking if the GPU itself is fully functional (die and VRAM), etc.

I never updated my VBIOS, how can I check my GPU is fully functional ?
I had green dots not on my TV but another screen so I assumed it was not the HDMI cables. I'll make sure tomorrow and try another 4K cable directly on my TV since this is where the problem mostly occurs.
Thanks for helping and the replies

 

[Newtonius]

I never updated my VBIOS, how can I check my GPU is fully functional ?

You can test your GPU through a stress tester like 'furmark' or 'unigine superposition', then compare the scores online to make sure everything is normal generally speaking. You can also use applications like MSI afterburner and HwInfo to monitor temps and power draw. This all gets a little technical, so try the other physical issues first before troubleshooting the GPU itself.
As for updating your VBIOS, it would be best if you create another thread about it so this one doesn't digress too much, updating VBIOS can cause major issues if done incorrectly. So again, save that for a last resort.

 

[Kurtz404]

You can test your GPU through a stress tester like 'furmark' or 'unigine superposition', then compare the scores online to make sure everything is normal generally speaking. You can also use applications like MSI afterburner and HwInfo to monitor temps and power draw. This all gets a little technical, so try the other physical issues first before troubleshooting the GPU itself.

I just tried to VRAM stress test my gpu (80% usage), and after 20 seconds I got 12500 errors, I kept the test running for 10 minutes and got errors every 20 seconds.

 

[Newtonius]

I just tried to VRAM stress test my gpu (80% usage), and after 20 seconds I got 12500 errors, I kept the test running for 10 minutes and got errors every 20 seconds.

Then there's your issue. Hopefully it's just a heating issue, check to see if your 2080 has sensors for the VRAM, you can use HwInfo to monitor sensor readings.

Free Download HWiNFO Sofware | Installer & Portable for Windows, DOS

Start to analyze your hardware right now! HWiNFO has available as an Installer and Portable version for Windows (32/64-bit) and Portable version for DOS.

www.hwinfo.com

When it launches, check the "sensors only" box and hit run. Scroll down all the way to your GPU section and see if there's any readings for your video memory.
If your GPU VRAM is also overclocked, then I would try downclocking them too, though I assume you didn't do that. If the vram is bad, then there's nothing anyone can do.
What test did you use by the way to test your vram?

 

[Kurtz404]

What test did you use by the way to test your vram?

I used OCCT VRAM standard test. I'm getting errors above 65%.
My GPU is indeed not overclocked.

If i'm looking for GPU Junction temperature I don't have this.
Here is what I got during a 4 minutes VRAM test (80%)
On GPU-Z, highest Memory temperature was 62.8°

https://imgur.com/lh1992V

View: https://imgur.com/lh1992V

 

[Newtonius]

I used OCCT VRAM standard test. I'm getting errors above 65%.
My GPU is indeed not overclocked.

If i'm looking for GPU Junction temperature I don't have this.
Here is what I got during a 4 minutes VRAM test (80%)
On GPU-Z, highest Memory temperature was 62.8°

https://imgur.com/lh1992V

View: https://imgur.com/lh1992V

That doesn't bode well, issue could be with your GPU VRAM then, and there's no fix for bad VRAM easily available, or at all really. Try dealing with the other issues and hopefully the solution is one of those other options I pitched. Updating VBIOS won't really fix that either.

 

[Kurtz404]

That doesn't bode well, issue could be with your GPU VRAM then, and there's no fix for bad VRAM easily available, or at all really. Try dealing with the other issues and hopefully the solution is one of those other options I pitched. Updating VBIOS won't really fix that either.

I guess I'm stuck with this issue then, i will downlock it and buy a new CG considering the price are lowering these days.
Thanks a lot for your help, much appreciated.