System with Nvidia card randomly freezing (and crashing)

There have been reported problems with the 470 driver, ttry see if you can get the latest or use an older driver.

I got an impression that it is an issue for those who have integrated GPU aside from NVidia. I never had an issue with .470 on my 2 desktops with NVidia GPU - none of them got an integrated graphics.

1 Like

Yes, a lot of notebook computers today, including gaming machines, have dual GPU's. My last gaming notebook had dual GPU's, and so does my new one. Whenever you have dual GPU's, it can invite problems.

Some issues can be solved with those by just going into the BIOS's, and telling your machine to use Nvidia online, or integrated only. Some machines like my new computer, have what is called a MUX switch, which is just a hardware level version of that switching.

Now, I wanted to tell you guys something, it is well known, if you have both GPU's enabled, the signal has to pass through the integrated, before it goes to your discreet GPU, causing a 10% difference, as in 10% reduction of total performance.

If you set your machine to Nvidia only, it bypasses the integrated GPU, and you gain a 10% increase in total gaming performance. However, something I noticed, when I tried to disable my integrated GPU in my computers BIOS, I noticed that my OS was acting graphically buggy.

When I got to the login screen, everything was just all wackydoodle and completely off. I discovered, that Linux wasn't going to let me completely disable the integrated GPU, because it obviously needed it, even if I had no intention on using it.

So I re-enabled the integrated GPU just to get Linux to straighten up and fly right. And thats when I later discovered the extension Nvidia Prime Indicator, which I use on the software level, to force use only the Nvidia GPU.

Its a win win. Linux is happy to see my integrated GPU, and I am happy, cause I get to use my Nvidia GPU only. I know, so weird, but Linux is weird people. What can I say? Anyways, thats my story.


1 Like

I will have to look into this, as I couldn't disable the integrated gpu without the same issues you described. I have it set to prioritize the nvidia card but couldn't force it to be used 100% of the time.

This is what I was using, who would've thought the tested ones are less reliable lmao. I'll switch to 495 I think

1 Like

Bad news, I have been continuing to have these crashes. Most commonly when loading in various games, but also when just using the desktop.

I'm not sure if this will give anyone a hint but the audio loops during the freeze up until the crash occurs. The last ~1/2 second of audio will loop.

I tried adding nvidia prime indicator but was unable to get it to work on my system. The drivers I am now running are the 495 proprietary drivers.

This sounds a lot like:

1 Like

Wish I could say I magically found a solution but unfortunately I have not. Sometimes a system update helps smooth things out for a day, but then I'm usually right back to the crashes. I checked the link @Aravisian provided to a similar issue but the solutions there didn't work for me either :sob: although they did make me think to mention that I am running a vpn on my system (windscribe)

I have tried all the available nvidia drivers, and the opensource ones too but they caused compatibility issues. The problem appears nonetheless with any of those settings.

I have tried prime indicator, but it does not prevent the crashes.

I am convinced that it is something to do with graphics as the crashes happen at an elevated rate when in games. (5-10 crashes per hour when in games)

This is a very frustrating experience as apart from these crashes I love zorin. :broken_heart:

Is there any chance to test your system with a different OS?

If the symptom disappear, it is a OS issue.
If the symptom persists, then it is a hardware issue.

If windows counts then the system is stable in win 10?
I'm hesitant to install more OS's as this is my main machine with all my data

1 Like

USB live session?
No installation require.

2 Likes

Well I have since tested with PopOS and the error happened there too. What would cause this instability hardware wise?

Cable connection, excess heat, and hardware failure.
Also slot fitting and insufficient power supply in case of desktop.

2 Likes

I never test anything on my main system as a Long ago learned lesson. For 50 buck US dollars (it's worth all the headaches that could come) you can get a decent machine to "play with" stuff till you learn the ropes... Just a FYI

Edit and if it's twice as much it is still a deal

Or this:

It can take all ARM version of Linux distros without any compatibility issue.

1 Like

Maybe santa-san will bring me one for a gift this tear... :thinking:

1 Like

i am going to be running rasdaemon to log what the possible hardware errors are and try to correct them. I also want to try an install of manjaro to investigate if arch plays nicely with my system

update
rasdaemon did not detect any hardware failures when the crashes were occuring. I also ran memtest86 and no issues were found with the memory. The gpu and all components are seated properly and my cables are all attached.
I could supply a syslog but I am not sure how to upload that here

These are the logs that I have recorded, unfortunately I don't know what any of them mean or how to solve them.

Kernel error.
Have you tried a Different Nvidia Driver?

You may try installing the Ubuntu mainline kernel:

sudo apt install linux-oem-20.04

Or you may need to upgrade to kernel 5.14. Due to an error on Ubuntu's part that would be very easy to fix, but they refuse to fix, everyone on Ubuntu 20.04 or Derivatives of 20.04 are stuck with a dependency error.
I have created a corrected Kernel for 5.14 that solves this error. Try the above suggestions and if they fail, I can PM you instructions on trying the 5.14 kernel.

2 Likes

Well I have tried other nvidia drivers and attempted what has been suggested so I'm gonna go for it with the mainline kernel and see what happens.
I guess if nothing else works it's good to know that it's a kernel error and leaving the ubuntu family is an option (solus or manjaro come to mind)

update
tried the mainline kernel and same issue occurred. Here are the logs this time:

So should I jump ship from ubuntu flavours altogether?

Can you try:

sudo apt install --reinstall zorin-os-desktop

And check your PM for a message from me on installing the 5.14 kernel.

You may need to jump the Gnome ship... Have you tested zorin OS Lite?

2 Likes