EmbarrassinglyParallel

The purpose of this project is to demonstrate different ways to parallelize a simple tasks that require no communication between the parallel processes, i.e. embarrassingly parallel tasks.

Slurm

The slurm directory contains a simple example of how to submit a job to a SLURM managed cluster to perform an embarrassingly parallel task.

• slurm/job_array.sbatch is a SLURM script that submits a job array to the cluster to parallelize tasks.
• slurm/two_process_example.sbatch is a SLURM script that submits a singular job to the cluster to parallelize tasks using two cores. This file also allows for one task to be started first.

Python

The python directory contains a simple example of how to parallelize a task using the multiprocessing and mpi4py modules in Python, and does not require a SLURM managed cluster.

• python/examples.py is a Python script shows how the below two files can be used to parallelize code.
• python/parmap.py is a Python script that demonstrates how to parallelize a task using the multiprocessing module, and has a parmap function that is similar to the map function in Python which can be easily imported into other codes. This script is designed to be run on a single node or a local environment.
• python/mpimap.py is a Python script that demonstrates how to parallelize a task using the mpi4py module. This script is designed to be run using the mpirun command (e.g. mpirun python examples.py), and can be used to parallelize a task across multiple nodes.