Run in HPC clusters

In supported clusters (currently Slurm and LSF), nnmk will also create a job.bash file. You can directly submit job.bash to the scheduler.

Cluster jobs also require a walltime property. If a job lasts longer than its requested walltime, nnodes will save and resubmit. You can disable auto resubmission by adding a property resubmit = False.

Cluster support

Nnodes has no strong connection with any specific job system, when running MPI tasks it just serves as a wrapper for commands like mpiexec, srun, jsrun, etc. So it is easy to add a new cluster configuration with the configuration file. Nnodes has built-in support for:

• Slurm

• LSF

• Local computer with multiprocessing