Fix a bug in a MPI python code

Job ID: 37367319

Budget: $10 – $30 USD

Hello,
the issue is on https://github.com/JohannesBuchner/autoemcee/issues/3 .

The problem is that I can't get to gather all the MPI chains. For the moment, once convergence with Gelamn-Rubin is reached, I get a final result ( the sentence "FINAL STATS") for each process, one after another. I would like to get only one final result ( FINAL STATS ) .

This looks like a problem of gathering on process rank 0.