Question Black Screen & GPU Fans to 100% when Gaming

[Floyd1319]

Hello, I recently ran into an issue when doing anything that is intensive on my GPU (gaming or Heaven benchmark with my testing) where my 2 monitors go black and my GPU fans go to 100% and sound like my PC is about to take off into space. The PC doesn't turn off on its own and the only way to be able to access anything on my PC again is by pressing the power button on my PC to restart it. All of my temperatures are well below what would be considered too hot. I have done so many things to troubleshoot this and I can't seem to find the issue. I've tried (not necessarily in this order):

• Unplugging all cables in my PC and plugging them back in (just in case I had a loose cable somewhere)
• Changing out my PCIE cable from PSU to GPU for a different one
• Cleaning the entire PC to remove all dust
• Removing and reinstalling the GPU after applying new thermal paste
• Removing the CPU cooler and reinstalling after applying new thermal paste
• Installing a brand new power supply (EVGA 1300W Gold PSU) and all the new power cords that came with it
• Using DDU to completely uninstall everything NVIDIA from my system and reinstalling the latest drivers
• Tried older drivers on my GPU
• Successfully ran Cinebench R23 multiple times so I don't think it is my CPU
• Updated BIOS
• Installed 2 PCIE connectors from PSU to GPU rather than 1 that had 2-6+2 connectors attached to it

I don't think I'm forgetting anything that I've tried, but I feel like I've done everything short of trying a new GPU (I don't have any extras or I would). If anybody could help me with this I'd appreciate it so much. I've spent a ridiculous amount of time troubleshooting this and searching for fixes and I just want the problem to be fixed so I can enjoy my hobby again.

PC Specs:

i9 9900k
ASUS ROG Strix NVIDIA GeForce RTX 2080 Ti
Samsung SSD 970 Evo Plus 1TB
EVGA 1300W Gold PSU
Gigabyte z390 Aorus Ultra Motherboard
G. Skill Trident Z 32GB (2-16GB) DDR4 3600

Again, any help is greatly appreciated!

 

[Aeacus]

Your MoBo has 3x PCI-E x16 slots. Have you tried different slots?

Also, while you did DDU and installed latest drivers, try DDU again but this time, version or two older than latest drivers.

When benching, have HWinfo64 open, with Logging enabled/started. It should save the system telemetry when you power-off after seeing black screen. And even if it doesn't, you'd see the telemetry just before screen goes black.

When screen is black, look how you can take screenshot with one key press (or combo press), with automatic image saving. This would tell if the screen is black even on the system side, or is the black image sent only to your monitors. (I'm using IrfanView for such automatic, single key press screenshots + image save.)

When screen is black, try the following combo: Win+Ctrl+Shift+B
This restarts GPU drivers and may bring image back, without the need of power-off/restart.

What Event Viewer says at the time when screen goes black? Namely "Windows Logs -> System" part.

 

[Floyd1319]

Your MoBo has 3x PCI-E x16 slots. Have you tried different slots?

Also, while you did DDU and installed latest drivers, try DDU again but this time, version or two older than latest drivers.

When benching, have HWinfo64 open, with Logging enabled/started. It should save the system telemetry when you power-off after seeing black screen. And even if it doesn't, you'd see the telemetry just before screen goes black.

When screen is black, look how you can take screenshot with one key press (or combo press), with automatic image saving. This would tell if the screen is black even on the system side, or is the black image sent only to your monitors. (I'm using IrfanView for such automatic, single key press screenshots + image save.)

When screen is black, try the following combo: Win+Ctrl+Shift+B
This restarts GPU drivers and may bring image back, without the need of power-off/restart.

What Event Viewer says at the time when screen goes black? Namely "Windows Logs -> System" part.

Thanks for the reply. I have checked Event Viewer, but the only errors that I'm getting are ones from turning the PC off improperly. I also tried Win+Ctrl+Shift+B and nothing happens. I haven't tried the other things that you mentioned, but I will tonight and get back to you.

 

[Floyd1319]

Your MoBo has 3x PCI-E x16 slots. Have you tried different slots?

Also, while you did DDU and installed latest drivers, try DDU again but this time, version or two older than latest drivers.

When benching, have HWinfo64 open, with Logging enabled/started. It should save the system telemetry when you power-off after seeing black screen. And even if it doesn't, you'd see the telemetry just before screen goes black.

When screen is black, look how you can take screenshot with one key press (or combo press), with automatic image saving. This would tell if the screen is black even on the system side, or is the black image sent only to your monitors. (I'm using IrfanView for such automatic, single key press screenshots + image save.)

When screen is black, try the following combo: Win+Ctrl+Shift+B
This restarts GPU drivers and may bring image back, without the need of power-off/restart.

What Event Viewer says at the time when screen goes black? Namely "Windows Logs -> System" part.

I tried to take a screenshot with IrfanView, but it didn't work when the screen was black. It seems like my computer won't allow any inputs when this happens because it worked just fine to take a screenshot before the screen went black. I again tried Win+Ctrl+Shift+B and this did nothing as well.

I installed HWinfo64 and have logs right before it happened, but I'm not 100% sure exactly what I should be looking at. I checked temperatures and a few other things and everything seemed normal enough, but I'm not sure if I should be checking something that I'm not because there are so many readings.

I tried a few different older driver versions after using DDU and none of them worked.

I also tried the other PCI-E x16 slots for my GPU and the problem persisted.

 

[Aeacus]

I installed HWinfo64 and have logs right before it happened, but I'm not 100% sure exactly what I should be looking at. I checked temperatures and a few other things and everything seemed normal enough, but I'm not sure if I should be checking something that I'm not because there are so many readings.

Yes, there are many readings. But what +12V rail readings are?
When you're at it, might also look +5V rail and 3VCC, 3VSB rails too.

--

Your RAM is running at 3600 Mhz, it wouldn't hurt to try running it at JEDEC speeds (2133/2400/2666 Mhz).

You also say that you have dual-monitor setup. Does the same issue arise when you use only 1 monitor?

Another option would be pulling GPU from system and hooking monitor to MoBo, to use CPU's iGPU. Now, iGPU isn't powerful but it is enough to get image on screen and perhaps even run lowest spec preset in Unigine (Heaven or Superposition), to test it. But since all fingers point towards your RTX 2080 Ti, i don't think you'll see any black screens when using iGPU.

Software wise, what "SFC /scannow" says?

(If you don't know what it is, or how to run it, further reading,
link: https://support.microsoft.com/en-us...em-files-79aa86cb-ca52-166a-92a3-966e85d4094e )

Once the scan is complete, a log file is created. It's path is written in the CMD. You can look it up, open it and read what exactly SFC /scannow found and fixed (quite an interesting read, at least, it was for me when i cleaned my system).

Now, i don't think that your issue would be software related. However, in a slight chance it may be, best option would be clean Win install. Since with it, you'd be formatting your OS drive and starting with a new and fresh OS, free of software issues (driver conflicts, bloatware, malware and the like).

--

Other than that above, you have little left to test. Those include:

• try with 2nd, known to work GPU in your system (to see if, perhaps, it's MoBo issue)
• try your GPU in another PC (to see if issue happens there as well)

Now, if your GPU produces the same issue on 2nd PC as well, your GPU is essentially toast. If you have warranty left, RMA your GPU.

 

[Floyd1319]

Yes, there are many readings. But what +12V rail readings are?
When you're at it, might also look +5V rail and 3VCC, 3VSB rails too.

--

Your RAM is running at 3600 Mhz, it wouldn't hurt to try running it at JEDEC speeds (2133/2400/2666 Mhz).

You also say that you have dual-monitor setup. Does the same issue arise when you use only 1 monitor?

Another option would be pulling GPU from system and hooking monitor to MoBo, to use CPU's iGPU. Now, iGPU isn't powerful but it is enough to get image on screen and perhaps even run lowest spec preset in Unigine (Heaven or Superposition), to test it. But since all fingers point towards your RTX 2080 Ti, i don't think you'll see any black screens when using iGPU.

Software wise, what "SFC /scannow" says?

(If you don't know what it is, or how to run it, further reading,
link: https://support.microsoft.com/en-us...em-files-79aa86cb-ca52-166a-92a3-966e85d4094e )

Once the scan is complete, a log file is created. It's path is written in the CMD. You can look it up, open it and read what exactly SFC /scannow found and fixed (quite an interesting read, at least, it was for me when i cleaned my system).

Now, i don't think that your issue would be software related. However, in a slight chance it may be, best option would be clean Win install. Since with it, you'd be formatting your OS drive and starting with a new and fresh OS, free of software issues (driver conflicts, bloatware, malware and the like).

--

Other than that above, you have little left to test. Those include:

• try with 2nd, known to work GPU in your system (to see if, perhaps, it's MoBo issue)
• try your GPU in another PC (to see if issue happens there as well)

Now, if your GPU produces the same issue on 2nd PC as well, your GPU is essentially toast. If you have warranty left, RMA your GPU.

All of the voltages seem fine to me.
12V: 11.952V - 12.096V
5V: 5.07V
3.3V: 3

Yes, there are many readings. But what +12V rail readings are?
When you're at it, might also look +5V rail and 3VCC, 3VSB rails too.

--

Your RAM is running at 3600 Mhz, it wouldn't hurt to try running it at JEDEC speeds (2133/2400/2666 Mhz).

You also say that you have dual-monitor setup. Does the same issue arise when you use only 1 monitor?

Another option would be pulling GPU from system and hooking monitor to MoBo, to use CPU's iGPU. Now, iGPU isn't powerful but it is enough to get image on screen and perhaps even run lowest spec preset in Unigine (Heaven or Superposition), to test it. But since all fingers point towards your RTX 2080 Ti, i don't think you'll see any black screens when using iGPU.

Software wise, what "SFC /scannow" says?

(If you don't know what it is, or how to run it, further reading,
link: https://support.microsoft.com/en-us...em-files-79aa86cb-ca52-166a-92a3-966e85d4094e )

Once the scan is complete, a log file is created. It's path is written in the CMD. You can look it up, open it and read what exactly SFC /scannow found and fixed (quite an interesting read, at least, it was for me when i cleaned my system).

Now, i don't think that your issue would be software related. However, in a slight chance it may be, best option would be clean Win install. Since with it, you'd be formatting your OS drive and starting with a new and fresh OS, free of software issues (driver conflicts, bloatware, malware and the like).

--

Other than that above, you have little left to test. Those include:

• try with 2nd, known to work GPU in your system (to see if, perhaps, it's MoBo issue)
• try your GPU in another PC (to see if issue happens there as well)

Now, if your GPU produces the same issue on 2nd PC as well, your GPU is essentially toast. If you have warranty left, RMA your GPU.

All of the voltages seem fine to me.
12V: 11.952V - 12.096V
5V: 5.07V
3.3V: 3.324V - 3.344V
3VCC: 3.072V
3VSB: 3.41V - 3.43V

It does still happen when only using 1 monitor.

I used SFC and it says that it did not find any integrity violations so it seems fine as well.

I tried turning slowing down my RAM to JEDEC speeds, but the problem persisted.

The only thing I didn't try was doing a fresh install of Windows, but I don't want to try that yet. I ordered a new graphics card and it is coming tomorrow. I'm going to try it in my system, and if it works, I know it is my graphics card. If it doesn't, I'm going to return the one that I bought and assume that it is the motherboard and order a new one and try that out. I'm half hoping it is the graphics card because I want the problem to be fixed, but I also don't want it to be the most expensive part in my system.

Thanks for all of your help!