MPICH question

I have a Docker image of my python program which calls into parallel version of meep (https://meep.readthedocs.io/en/latest/) which is compiled against the MPICH (pymeep package in conda-forge). I'd like to run it on the K8s using MPI-Operator. I see 3 possible paths:

1. Extend MPI-Operator to support MPICH
2. Recompile Parallel Meep to use OpenMPI
3. This whole things won't work either way because of shenanigans of havin MPI enabled app inside conda environemnt in docker which is not supported.

Any advice which is better? Will it even work or am I doomed and I'll hit point 3 above?

Start by reviewing MPI-Operator's documented support for MPI implementations and the linked Meep/pymeep container setup, then compare the three paths described in the issue. Done would be a maintainer-confirmed recommendation or a clearly scoped change for running this workload on Kubernetes.

Written by the indexing model from the issue text.