I am trying to run my analysis on a super computing center and I am facing some difficulties. I wonder if anyone has experience running openseespy in HPC. Thanks!
It seems that openseespy does not support Parallel Scheduler. Although parallel computation is possible in one linux system, it cannot be used in HPC Cluster (multiple nodes).
On the openssespy site, there is no information about Parallel Scheduler support.
However, it may be possible because it uses the mpiexec command. However, this will only be possible with support from the openssespy library.
I tested it several times, but it didn't work.
I have installed miniconda on the cluster now.
When I tested the sample code, hello.py, it doesn't seem to be supported.
When I test it, I get error. (not support parallel processor)
Thank you for the quick response!
This is what I got.
When it comes to parallel calculations, there are two cases.
First case, how to parallel calculate one node CPU
The second case, which calculates parallel calculations of multiple nodes and each node CPU.
The reference video is the first case.
Only the second case is available at the Supercomputing Center using the SGE scheduler. Therefore, OpenSeesPy should support the SGE scheduler.
Please check this part.
I want to know if Openseespy support the SGE scheduler.
Thank you!
,