@encadyma suggested that we could run a render farm server, like flamenco. In addition, some others have suggested running map/reduce on HPC.
If we do any of these, we'll need to use some cgroups-like resource limit mechanism to make sure that they don't step on SLURM or vice versa (or we run them on different nodes).
For the render farm idea, we might consider reaching out to UCBUGG. They are a decal who teach graphics to students, and when I took it, I believe they were short on render time.
@encadyma suggested that we could run a render farm server, like flamenco. In addition, some others have suggested running map/reduce on HPC.
If we do any of these, we'll need to use some cgroups-like resource limit mechanism to make sure that they don't step on SLURM or vice versa (or we run them on different nodes).