GPUs not available to select for multi jobs

Hello,

I’ve started to use CryoSparc and I have a problem when running multi-GPU jobs. For some reason, I cannot chose which GPUs to use for the job:

The job runs but automatically assigns GPUs starting from 0.

Can anyone help me with this problem?

Thanks in advance.

Hi @LTP,

Based on the screenshot you provided, it looks like cryoSPARC wasn’t able to get information about the GPUs on this node - this happens every time cryoSPARC is started. Could you please provide the following info:

  1. Is cryoSPARC installed in a single workstation setup (the master and worker components are installed on the same machine)?
  2. Try restarting cryoSPARC: cryosparcm restart. When cryoSPARC is running, please run cryosparcm log command_core and reply to this post with the output

Thanks,
Suhail

Hi @sdawood,

Thanks for your reply.

  1. Yes.
  2. Probably it is a very dumb reason, but I cannot run any cryosparcm command (it doesn’t recognize the command). I just installed the software following the instruction guide and use the web-based GUI. Should I do something to activate it in the terminal (I’m logged in via ssh to a computer with the GPUs I use to process).

Thanks for your help.

Hi @LTP,

No problem, we’re here to help! The most likely reason you aren’t able to run the cryosparcm command is because it isn’t part of your $PATH variable. You can still run the command by first navigating to the cryosparc_master folder. Then, just run ./bin/cryosparcm <command> - so to restart, just:

./bin/cryosparcm restart

After cryoSPARC starts again, please run and post the output of the following command:

./bin/cryosparcm log command_core

Hi @LTP,

Now that they show in the UI, you don’t need to use the command line anymore. It looks like the restart solved the issue!

- Suhail