I have two computers with near identical specs (only thing different is 5820k vs 5960x CPU), both of them have a GTX 1070 with NVIDIA driver 470.82.00. However, one of them has the screen randomly going black for 1-3 seconds - the system is responsive, it’s just the screen goes black then goes back to normal as if nothing happened. Both of them have a VG248QE primary monitor hooked up via DP and a PA246Q secondary monitor hooked up via HDMI. I’ve tried 3 different HDMI cables and also swapped out the DP cable for the primary monitor for good measure as well but the issue still persists.
There isn’t really a sign in the logs that the monitor was off and on again. Did you try swapping monitors with your other system to check whether it’s the monitor itself that’s failing?
Thank you for responding, my other thread never got a response so I was wondering if I was doing something wrong (although that problem had a fairly good workaround at least).
Unfortunately the other system is two hours away (with the known good monitors), but, your idea is certainly worth checking early next year when I have a chance.
I was mostly just wondering if like you said there’s anything in the logs to indicate that something is going wrong. I captured the logs after the monitor went off/black twice already and was actually already going black during the log capture, so I thought there was a high chance of the logs picking up something.
Are there any other logs I could post other than the NVIDIA generated logs and dmesg?
No other logs needed. Whenever something changes in regard to monitorsettings and connection, this gets logged to Xorg.log.
You could run
tail -f /var/log/Xorg.0.log
(maybe using sudo)
in a terminal windows and check if something gets logged when the monitor goes black.
BTW, unrelated to your issue but you should check which USB device is broken and gets reset all the time, flooding dmesg.
I will post a capture of Xorg logs next time I see this issue happen (I’m on a fresh reboot at the moment).
For the USB issue it’s more of a known-issue sort of thing with my platform, I’m on X99 which has had quite some… interesting USB problems to say the least lol. All my USB devices are functioning properly thankfully, but yeah I have to put up with some log spam lol
So as far as I can tell, the black screen bug didn’t happen today. I don’t necessarily think it’s solved but I do think it’s intermittent so this might be hard to track down. I will have to keep tail -f /var/log/Xorg.0.log in mind especially on days where I notice the black screen issue happening often.
In my Xorg.0.log I noticed these messages about monitors connecting and disconnecting, however, it’s saying that both monitors are disconnecting and reconnecting, and yet I’m not even noticing any black screen issues today so I think the logs are likely unrelated. Here’s a picture of the logs with the disconnecting I’m talking about: https://i.imgur.com/nYVZ8KV.png
I really think your monitor has a slight damage since when it went black, nothing showed up in the logs. This might also be a compositor bug but chances are really low.
My second screen (Dell S2721DGFA) has been showing exactly the same behavior for a few days now and the screen itself is perfectly fine (flawless on other OS).
But i ran some apt updates and i can’t remember if the driver itself was also updated, at the moment it has 470.82 on it.
The screen is black (only on second screen) for 2-3 seconds after opening a new Instance of Mozilla.
I won’t have access to the computer with the issue in question until probably March or April of next year, so this will probably be my last reply on this topic for a while, but I’ll probably be back eventually unless NVIDIA manages to fix it with a driver update in the mean time (in which case I’ll still probably be back).
I think my next steps for this will be factory resetting my 1200p monitor, trying a third different HDMI cable (a known good working one that I use on my main PC) and then also trying out a different monitor to see if it still happens. Although if it doesn’t happen with a different cheaper 1080p monitor, it could still be a GPU driver issue or GPU hardware issue not necessarily the monitor itself since my 1200p monitor would be higher bandwidth than a spare 1080p monitor I have access to.
I still won’t have access to the computer in question until April but I am absolutely not going to be using stock clocks, that’s not an option. I am not going to be giving up 30% of my CPU clock speed. My CPU overclock is entirely stable (can run 10+ hours of prime95 no problem) and my CPU has nothing to do with my display because it’s not a laptop, it’s a desktop.
It seems like this is now solved. I updated the driver on the computer in question to 470.103.01 and the issue went away entirely. After that, I upgraded my GTX 1070 to an RTX 3070 Ti and the issue still did not happen. Later on, I upgraded the driver to 510.73.05 and the issue still did not happen.