Running ipyparallel cluster on multiple nodes using MPI

I am trying to start an ipyparallel cluster using MPI.

In ipcluster_config there are the following lines, modified as such:

c.MPILauncher.mpi_cmd = ['mpiexec'] c.MPIControllerLauncher.controller_args = ['--ip=*'] c.MPILauncher.mpi_args = ["-machinefile", "~/mpi_hosts"] 

ipcontroller_config.py is configured as such:

 c.HubFactory.engine_ip = '*' c.HubFactory.ip = '*' c.HubFactory.client_ip = '*' 

However, when I start the cluster using the ipcluster start --profile mpi -n 2 it does not work with the following message

 Engines shutdown early, they probably failed to connect. You can set this by adding "--ip='*'" to your ControllerLauncher.controller_args 

Not sure how to debug further.

+5
source share

Source: https://habr.com/ru/post/1273291/


All Articles