I have a random ocasional problem, sometimes it does not rear its ugly head for a week and then sometimes I crash a few times in a day.

I will use my current game as a example, when playing FFXI online mmorpg every once and a great while when i zone in or turn a certain way my screen locks up and becomes puke green one color the system is still moving my music is still playing I hit ctrl+alt+escape to kill the puke green program on screen 0, then that screen goes black, and about 10 seconds later as im trying to fiddle with control pannels on my second screen the system freezes, it also still freezes even if im super fast and end the x server right as it goes puke green.

This also has happened on Quake 3, and on Counter-strike source. So it makes me think its a nvidia related problem most likely but does not seem to be affected by changing nvidia driver versions so i reverted to the version that works best overall for the older games I play.

Any idea how one might start to debug and find the culprit that freezes me?

Somehow I doubt there will be a solution but its been bugging me for almost a year now and still happens with nvidia current drivers as well, and figured id finally sit down and post the problem as well as possibly see if anyone else has had this happen.

System spec
HP Z840 Workstation Dual Xeon 2.6Ghz 12 core v2(48 cores)
64gig ddr4 ecc ram
Geforce 980GTX 4GB
2TB NVME on a pci-e card root drive /

Slackware 15.0
Kernel 5.15.28 LTS
Nvidia 470.86 commercial driver
KDE Plasma 5

Sounds very much like a GPU issue, related to either cooling or power supply.

Have you checked the log files.

It sometimes happens in a system freeze that the freeze is so quick nothing gets logged, but it's worth a look. I'd start with /var/log/syslog and /var/log/Xorg.log.

If you do this immediately after a freeze and restart, the tail command can be quite helpful. See man tail for more.

I checked didnt see much. I have usually 5-10 seconds before full pc freeze after the screen dies, sometimes its puke green sometimes its brownish but its always one solid color the second it dies, and its only on the main screen during a 3d game. There is usually enough time for me to hit the skull and crossbones shortcut and kill the game window. Then as im trying to open nvidia control pannel to see if i can enable/disable that screen or something to toggle it it hard locks and needs to be powered off.

Cleared all logs and going to start from scratch and see what pops up right when it dies and write down the moment it crashes. It may be today, it may be later this week I will play FFXI religeously which should eventually trigger it.

I hope its not a hardware issue, I keep my pc clean and dusted out every 2-3 months, and it almost never shuts down once booted. System Temperatures are great.

My longest uptime was around 90 days then the power went out. Heck I still have a 10,000 rpm hard drive with 10 years spin time in this pc thats from my last pc which I still use as a scratch drive. Though those uptimes were usually when I wasnt gaming much.

Even heavy code compiling using -j48 for hours at a time doesnt bother this computer it seems just games which hints to me something GPU/GPU Driver related maybee even with dx8 api's that old games use or something to do with wine or proton. I built the unreal editor source code in 30 minutes no problem. Kernel build is only like 5-10 mins~. Since im using slackware builds I compile alot of stuff.

When I was doing some CPU AI image generation and stuff didnt have any issues then either with 100% cpu and 62gb ram used~. So I assume the enterprise 1400W PSU/Xeon cpus/motherboard in this beast is good too.

Memtest results are also clean and am using enterprise grade ram sticks 4GB*16 slots.

Once it happens again I will post one more time my findings. And probably some more quesitons.