GPU dies few moments after cryoSPARC is started

Install, Configure and Update
1

One of our stand-alone workstation’s GPU dies after cryoSPARC is started (using cryosparcm start). Sometimes right away, sometimes a few minutes after. This happens even without starting any cryosparc jobs. It causes the screen to black out as our GPU is needed for the display. I can ssh into the workstation remotely on command line; however, typing “nvidia-smi” completely freezes the terminal without any output.

Has anybody experienced the same problem and can give some suggestions on how to proceed?

I’m thinking of completely uninstalling cryosparc and re-installing it, but there doesn’t seem to be a set of clear instructions for doing that? Is it possible for someone to tell me or point me to resources for uninstalling cryosparc completely?

Things I’ve tried so far to no avail:

  • update cryoSPARC to the latest version: v4.7.1
  • update Nvidia driver to the latest version: 570.133.07

About our workstation

  • single workstation
  • Ubuntu 20.04.6
  • NVIDIA GeForce RTX 4090 x 2
  • NVIDIA driver version: 570.133.07
  • CUDA Version: 12.8
  • cryoSPARC version: v4.7.1

output of “name -a && free -g”

Linux sn4622117815 5.4.0-216-generic #236-Ubuntu SMP Fri Apr 11 19:53:21 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
total used free shared buff/cache available
Mem: 251 1 248 0 0 247
Swap: 17 0 17

Assistance is greatly appreciated!

2

That sounds like a hardware fault, more than a software issue.

Uninstalling CryoSPARC is easy: detach all projects and delete the cryosparc_master, cryosparc_worker and cryosparc_database directories. But I doubt it will solve the issue.

Try running CryoSPARC on the GPU not attached to the monitor, see if the same thing happens (in which case it’s the driver panicking) and check dmesg for nVidia related errors regardless of whether or not the monitor cuts out.

Also check in the job log of failed jobs.