Is there any way to visualize results of Python script using a plotting library such as matplotlib or seaborn from the cluster itself? If not, what would be a good work around to this workflow?
It might not answer your question directly but a workaround might be to use OnDemand interfaces. You can visualize python plots via Jupyter Notebook interfaces. Maybe you can look to run your workload on clusters in the ACCESS system that provide such interfaces.
Jupyter Notebooks should be available on clusters in the ACCESS ecosystem that have OnDemand web-based HPC interfaces. An example is Purdue’s Anvil system.
Which cluster are you currently using in ACCESS? You might be able to find relevant Jupyter-usage articles in the respective knowledge bases. The following are examples of guides to Jupyter resources on a few HPC systems in the ACCESS ecosystem.
- TACC Frontera: Jupyter Sessions - TACC Frontera User Guide
- PSC Bridges-2: Jupyter notebook instructions | PSC
-
Expanse SDSC:
– Training - SDSC Webinar: Run your Jupyter Notebooks anywhere: Scaling up your Projects from your Laptop
– Documentation - Running Jupyter Notebook on SDSC HPC System
Adding to @Fahad_Khan if your institution doesn’t allow this capability you could ask the HPC team if they would consider setting it up. The quickest way to do this is using a jupyterhub container but your system administrator may poke a hole in the firewall to do this. This is why you talk to your HPC team