Computer specs? PSU issue potentially?

If you have another certified DP cable then try that first or if you have an HDMI try that instead just to see if your cable is causing this. 

If you have a spare PSU with high power spec, it does not hurt to try with different PSU. A lot of time,  the issue can be fixed by changing to a high power PSU.

I have the same problem with 3080 Ti. No response on the ticket I submitted yet. Please let me know if you guys find a fix. 
 
Specs: ROG Dark Hero, Seasonic Prime TX-850, 5950X, 48GB DDR4, 4TB SSD

I have the same problem with 3080Ti FTW3. I have RMA'd the card but also with the replaced card I have the same issue. 
I have also tried to change the PSU to a 850w from Cooler Master, but still the same issue.
 
I suspect it's either a driver issue or a FW issue.
 
I can play League of Legends without many issues, but I found a somewhat reliable way to trigger the issue that is running OCCT with shader complexity 4 and have Twitch open on another screen.
Last two times I left it running, when away from the PC for few minutes, came back with one display off and the drivers had crashed.
 
Another game I'd like to play is Prepar3D - in that one instead it's even happening with higher frequency: I cannot play for more than 10-15 minutes.
 
I am somewhat relieved to see that I am not the only one.

Steps I tried (answering to Sajin):
 
#1 RMA'd the card. I have two cards that exhibit the same issue
#2 Covered by 1)
#3 They are certified and worked for more than a year with other cards
#4 Replaced with a (bigger) 850W from Cooler Master
#5 Tried right now with Gen4 - still happening
 
For completeness of information, after the black screen only one display comes back up alive. The other one stays dark (unless I reboot the system).
In the Event Viewer I always get the same error:
 

The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event:
\Device\Video3
CMDre 00000001 00000220 00040340 00000004 00800000
The message resource is present but the message was not found in the message table

Sajin

franz86
Steps I tried (answering to Sajin):
 
#1 RMA'd the card. I have two cards that exhibit the same issue
#2 Covered by 1)
#3 They are certified and worked for more than a year with other cards
#4 Replaced with a (bigger) 850W from Cooler Master
#5 Tried right now with Gen4 - still happening
 
For completeness of information, after the black screen only one display comes back up alive. The other one stays dark (unless I reboot the system).
In the Event Viewer I always get the same error:
 

The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event:
\Device\Video3
CMDre 00000001 00000220 00040340 00000004 00800000
The message resource is present but the message was not found in the message table

Test the card in another known good working system. Just because you rma’d it doesn’t mean you got a good card back.

That's what support also told me but unfortunately I cannot. Does this also mean that other people here have broken cards? What are the odds anyway? And both cards are damaged in the exact same way?

I've tested this all day today (at least 6+ hours PC was on), and after a lot of research and a hunch I had... I may have fixed it. Granted, my 3070 Ti is a 290 watt card, but I was running a single cable with the dual 8-pin splitters. Sure it worked fine for the first couple months I had the card, but then.... this started happening. It's definitely hardware, because it still happened when I tried attempting to reinstall Windows 10. The INSTALLER BLACKED OUT.
 
Use dedicated PCI-E cables for each plug on the card. Do NOT use the y-splitter pig tails.
 
I've read so many stories of 1080 Ti's, 2080's, 3080's, etc running into issues like this with the y-splitter cables. Some fail quickly... others could take a year. As soon as people ran each 8-pin (or even 6-pin) as one dedicated cable each to the PSU, problems resolved. Even Jays Two Cents had a rant about this y-splitter bad practice just a couple months ago.
 
So yes, it could be PSU... it could be GPU. The answer could be both, because the power spikes the GPU draws may be undercutting the PSU's ability to supply all the power through those Y-splitters. The disconnect sounds the PC makes at the moment the monitors go black is a driver failure, EVEN OUTSIDE THE OS DURING THE INSTALLER PROCESS. This reply is mainly for higher watt cards. I never had this issue on a 1070 or 2070 Super.
 
Give it a try. I'm still monitoring it, but I went from 6 monitor black outs in 3 hours, then did this, it's been perfect for another 10.
 

Notice I'm using separate cables and disregarding the splitter sections. (the white cables are extensions)

I had this issue start this week with a 3080 TI FTW3 in PC with spec's below. I had no issues with the card since it replaced a 3070 XC3 ultra July 5th and had no issues with the PC when using the 3070. I was able to browse the web without issue, but any load and I would get black screens and see no data from the GPU to my in case monitors (for example display in PC case would show GPU temp going to 0deg). The PC would still run and I could hear audio (i.e. me crashing in the race when the screen went black). At the same time the BSOD started I was also hearing random disconnect sounds when I would login to windows. I tried different DP and HDMI cables but that did not solve the issue. 
 
I did a clean windows install and the disconnect sound went away. I still had BSOD issues and couldn't complete 3dmark runs so I then switched to the OC bios to test and the unit worked without issue so I then switched back to normal bios and reinstalled all drivers and didn't have an issue after that. I did upgrade my PSU from a 750W corsair to a 1000W corsair to eliminate that as a potential cause. I was not using daisy chains before or after. I also used a different surge protector in an abundance of caution. After 3 hours of use last night the issue didn't reappear. 
 
PC:
 ASUS B550 Gaming F-wifi
Ryzen 7 5800X (testing with & without PBO on)
16GB (2x8GB) G.Skill 3600mhz (my next upgrade path)
EVGA Nvidia RTX 3080 Ti FTW3 Ultra
NZXT Z63 280mm AIO
Samsung 970 Evo plus 1TB M.2 with Samsung 860 1TB SSD
Corsair RM750W replaced by Corsair RM1000X 
Fractal Meshify S2
3 x AOC C32G2S 2k 165hz using DP
6 x 140MM case fans

I was able to fix the issue by disabling GSync. As soon as I changed that to Fixed Refresh the whole system became mega stable. Never had any issues.
 
Does this ring a bell? I read that it's a known problem with Nvidia drivers and people having dual-monitors setup.
 

I am glad I found this thread. I was planning on using the Y splitter cable for my GPU that I am about to get, but after reading this I will make sure I use 2 cables instead. Good info!

superflydgreat
I am glad I found this thread. I was planning on using the Y splitter cable for my GPU that I am about to get, but after reading this I will make sure I use 2 cables instead. Good info!

If the card has 3 sockets, then 2 cables are enough. You put the first one straight in, the second one must be Y splitter /daisy chain so that you can use it for socket #2 and #3.