I have a Gygabyte MW51-HP0-00 motherboard with an Xeon W class CPU (Intel® Xeon® W-2145 CPU @ 3.70GHz) and an NVIDIA GTX 1080Ti. the video card drives 3 32" 4K Dell U3219Q monitors. The monitors are running in full 4K@60Hz. The linux distro is Fedora 30.
My X11 server crashed today. The symptons were the following.
I powered up my system, logged into the gnome desktop. I then started up firefox and saw a few videos on utube. I full screen the videos and all is working well. I then fire up a 2nd firefox window and log into the nytimes web site. when I do so, I notice that all video is slowing down. the utube video I’m watching begins to stutter, but the audio continues OK. The responsiveness of the nytimes firefox window is also very sluggish. It’s hard to scroll up and down through the web site. A lot of animated ads are running on the nytimes web site. I supposed that it’s the nytimes website which causes the problem.
I kill the nytimes firefox window and this brings my system back to normal. I continue to watch a few more youtube videos when all three monitors go black. I log into my system from my laptop. (I ssh in) and to a ‘top’ and see the Xorg process running at 100% cpu.
I run the nvidia bug report tool and the reboot. It takes too long to reboot, so I just leave my desktop in the reboot mode and go and have breakfast. When I get back, (a few hours later) the system had rebooted.
I toss this bug report out there. I hope someone can diagnose it.
nvidia-bug-report.log.gz (1.15 MB)
You’re running into an XID 79, most often caused by overheating or insufficient/faulty power supply.
I’ll swap out the power supply. Thanks.
I did a little total power draw testing on my system. I have a device which measures the power on a power cable. I have it plugged into my UPS which powers my system. The UPS powers all 3 monitors, the PC and a few misc items like my network switch, phone etc. When I power off the system, the total wattage is about 20. Basically the monitors are in stand by mode plus network switch and phone. When the monitors are powered on, the draw pumps up to 150W. when I boot up my system, I log in and its basically idle, the power draw is about 220W. When I fire up my windows 10 VM, the power will peak at 350W. When, the vm goes idle, the power drops down to about 250W. When I fire up firefox and surf to the nytimes web site, with lots of animated adds and inlaid videos, and I fire up another fixefox window which I have a utube video running in full screen, the power then goes up to between 300 and 350W.
As for the temperature of the video card, it nevers seems to get above 54-55C.
I have a 600W powersupply which should be enough to not cause power issues with the video card when running under these conditions.
What I’m wondering is if there is something else that glitched, which caused the video card to run hard, and then draw lots of power which then caused the XID 79.
I’m still going to swap out the powersupply, but I think moving forward it would be great if there was a howto on enabling debug mode on the video card so that there is extra output in the nvidia bug report which will help you find in more detailed what happens when the system glitches out. I’ve posted problems with my RTX 2080 Ti, and I’d like to run more tests and send feed back to you. Once I have my new power supply in I would like to turn on debug mode and proceed with testing.
With XID 79 there’s no further debug info available since it’s kind of an emegency shutdown to prevent hardware damage. Besides psu/temperature there are also other reasons possible, to a very lower degree, like bad slot/connection, gpu board failure, mainboard/bios problems, ripple on power cords, etc.