Dear all,
I have CS installed on my own workstation. In oder to check the performance with other GPU’s, I’d like to use Cryosparc on a cluster of GPU’s that I have access to.
Is it only possible if I am actually able to install CS on the worker node? Or can I somehow different access GPU’s from a cluster?
Best
Dario
cryosparc_worker software must either be installed on or shared with the putative GPU worker nodes and the cluster connected to the CryoSPARC instance.
Ok great, thanks. So if I am not allowed or if it is not possible to install any software on the GPU node directly, how does it work?
CryoSPARC can and should be installed under a non-privileged Linux account.
- Do you have a Linux account on the cluster?
- Do you have space and permission to install software under that Linux account?
Caution: Installation of the software is only one of several pre-requisites. Other prerequisites are
- shared project storage between master and worker(s)
- a common Linux account between master and worker(s)
Alternatively, you could install a self-contained single workstation-type CryoSPARC instance on the cluster, subject to positive answers to the questions above.