Nvidia driver 525.85.05 regularly crash on rtx4070 or rtx4080 / ubuntu 22.04

[quentin.deyna]

Hi everyone,

I regularly get a crash (like every 3 or 4 days) with nvidia driver 525.85.05
We work on omniveres/isaac-sim and it’s the recommended driver.
We install it through th .sh installer.

Here is our config :
CPU: 16-core (8-mt/8-st) 13th Gen Intel Core i7-13700K (-MST AMCP-)
speed/min/max: 835/800/5300:5400:4200 MHz Kernel: 6.5.0-14-generic x86_64 Up: 1h 11m
Mem: 4043.6/64063.8 MiB (6.3%) Storage: 1.82 TiB (7.9% used) Procs: 474 Shell: Bash inxi: 3.3.13

Here is the bug report
nvidia-bug-report.log.gz (5.0 MB)

My “routine” when I have this crash is to reinstall the driver but sometimes I get the “warning: the compiler differs from the one used to build the kernel” error and can’t reinstall the driver.

here is the install log :
nvidia-installer.log (144.9 KB)

Thanks for your help and have a good day !

[quentin.deyna]

ok so I manage to install the 525.147.05 .run version.

It seems that I had a auto kernel update this morning and this might be the reason the 525.85.05 driver crashed and failed to compile after a clean nvidia driver reinstall.

I’m not sure if the 525.147 will work well with isaac-sim since it’s not the one recommended but isaac start and work correctly for now.

As a precaution, I hold the kernel update with :
list the kernel packages installed :

sudo apt list linux-$(uname -r) --installed

and then hold each of them :

sudo apt-mark hold package_name

finger crossed!