It would be useful to be able to specify which GPU to use for a particular job when executing the job. Example situation - I have four compute GPUs, which I use for a mix of relion and cryosparc jobs. Right now I am running a long relion job taking three GPUs, but I would like to be able to queue up cryosparc jobs on the remaining GPU. Currently there is no way to do this without globally changing which GPUs cryosparc is allowed to use on the command line, which is a little cumbersome to do each time.
Thanks for bringing this up. We’re currently in the process of developing a new scheduler that will allow you to run jobs on specific GPU’s on specific nodes and also have the option of running multiple jobs on a single GPU to support non-process exclusive GPU’s.
Oh that’s great Stephan! I’ll look forward to it.
Will it be available throught the command line, that would be great.
Yes, overriding the scheduler and queuing a job directly to a GPU will be available via the CLI.
Thanks! that is excellent.
Let me tell you, that many of our CNB users (most of them use Scipion) now they are using Cryosparc through its plugin, and they are very happy with the results and performance.
This has been done as of v2.12.