Xid 79: GPU has fallen off the bus on 3050ti laptop

Graphics / Linux Linux
1

Hi. As the title says, I have problems with my gpu falling off the bus for seemingly no reason. The time it takes for this to happen is also random. Sometimes I can go hours without an error (even while playing graphically demanding games), and other times it happens as soon as I log into my user (or even before that). This problem has been happening for around 3 weeks now, and during that time I have:

-Tried moving from windows to linux to see if something driver-related was causing the issue, didn’t work.
-Tried limiting gpu clock speed, didn’t work.
-Tried improving cooling to my laptop, didn’t work (and even then, it kept crashing even at low temperatures).
-Tried activating persistenced mode (since I kept seeing that as a fix) through “sudo nvidia-smi -pm 1”.
-Downgraded and outright reinstalled my linux distro, didn’t work.
-Updated my BIOS, didn’t work.

Screenshot_20240920_023740
Screenshot_20240920_023740950×470 99.3 KB

Here is an image of journalctl after the error.

I have also attached the nvidia bug report just in case.
nvidia-bug-report.log (2.3 MB)

I am at a loss as to what the problem could be. Any help would be greatly appreciated. Thank you.

2

I have been having this exact error multiple times a day on my 4090, for the last few weeks. Linux kernel 6.10

Also updated my bios, and enabled persistence mode. I just enabled persistence like 15 minutes ago, and it hasn’t happened yet.

This doesn’t happen while gaming, just when the pc is idling at the desktop.

3

So it’s happening under Windows as well?

4

Yup, event viewer after a crash showed that nvlddmkm.sys was having problems, and after a full reinstall it didn’t go away so i switched to linux.

5

Persistenced didn’t work as a solution unfortunately.

6

Hmm something is probably wrong with with your hardware then if you experience issues in windows and Linux.

7

I second that, it’s broken, please check your warranty status.

8

It seems to have worked for me fortunately. I don’t know about windows these days, but yeah, you might have a hardware issue.

The only suggestion I have is that the way you are launching persistenced needs to be done on every boot. Maybe you did not know this.

You can add that command to something that will autostart it like rc.local

Or you can use
“systemctl enable nvidia-persistenced.service”
“systemctl start nvidia-persistenced.service”

and it will run itself when you start up.