Intermittent complete loss of video

Wyrdsmyth

I've got an off-brand Radeon RX 580 in my custom desktop computer that has started acting up over the past few months. I'm trying to make a determination on whether the issue is hardware, and if it is whether the issue is at the motherboard or the video card. I've been running the same installation of Solus (with every update) since 2021, if that matters. I am using the default Radeon driver, but previously I had an nVidia card installed and was using the proprietary drivers for that card.

Starting a couple months ago, intermittently, and almost always while playing a video game of some kind, the video output stops completely. I can usually continue to hear audio and for several seconds it appears that the system is continuing to operate sans video, but then there are indicators that the system has suffered a hard-lock of some kind. Either the numlock key doesn't respond to key presses (one of my typical tests I use) or the audio begins to rapid loop repeat. In all cases the only option I have to resolve the issue is to force the system to power off.

The issue is more likely to occur if the computer has already been running for several hours. If I shut down the computer, turn disable the power supply and wait about 20 seconds before starting it back up, the issue is less likely to occur, so I'm wondering if its possible that there's a memory related issue that could be part of the problem. One thing I haven't tried yet is a memory test so I'll give that a shot, especially since I recently upgraded from 16GB to 48GB with a 32GB kit (2x16).

Is there a log I can review that would help me narrow down the cause of the issue? I've experienced this video failure in both 3d and 2d games including Cyberpunk 2077 (Lutris), The Long Dark (Steam) and Death Must Die (Steam). I'm just not sure where to look, and perusing the boards I'm assuming I'm the only one having this problem, which means it's probably hardware.

Thank you for any suggestions!

infinitymdm

I have a friend who is experiencing similar behavior on Ubuntu. He came to the conclusion that his GPU was reaching temperatures that were unsafe and triggered an emergency shutoff. (I'm not entirely convinced that he's correct there, and I've offered to redo the thermal paste on his GPU, but he has yet to take me up on that offer 😁)

From my own experience I've seen a very similar issue when overclocking RAM too high. This was using XMP/DOCP profiles that the memory was (supposedly) rated for, but I guess you don't always win the silicon lottery. You can try disabling those in BIOS and seeing if you still have problems.

If you want to be sure it's hardware, I would try booting up a live image of some other distro. Load an application that puts a huge load on the GPU, like a blender demo render or something, then wait to see what happens. If you get the same behavior, you can be pretty sure it's a hardware issue.

Citfta

I agree it sound like a hardware problem. I have had similar issues when my CPU fan was going bad. It didn't affect the video. It just slowed the whole system down and sometimes locked up. I would suggest checking to make sure all fans are clean and working. There are also programs you can install to monitor the temps of the CPU and GPU. They would probably give you a good idea what is going on.

Axios

This is a small program prob run from live ok I havent used it but looked at seems like it would work
also shows pwr supply voltages.
https://www.ocbase.com/download
It can monitor to dunno if that would be of help

TraceyC

I agree that this sounds like a hardware problem. The audio looping after a bit (after its buffer is gone through) says it could be GPU, CPU or RAM. (If it was the power supply I'd expect different symptoms).

In addition to the good advice already in this thread, I'd like to add just one thing. Take a logical troubleshooting approach of testing one piece of hardware at a time to see if you can narrow down the root cause. You've already mentioned testing the RAM, that's always a good first step. As you go through the good ideas here, take it one component at a time.

Good luck!