Quest: MPI and mpi4py

advanced HPC


Overview

Do you want to learn how to parallelize your code to use the cores or memory of more than one node in a cluster? Did you know you can do this without Fortran or C++? Message Passing Interface (MPI) is a standard that allows researchers to leverage the computing resources of clusters like Quest to execute compute-intensive and memory-intensive code across nodes. Mpi4py provides an implementation of MPI for Python. We will be reviewing the basic ideas behind MPI, such as communication and distributed memory, and then work through Python examples that demonstrate how to incorporate the mpi4py interface in your work.

Prerequisites

  • Knowledge of the Linux command line, at least at the level of Command Line: Introduction
  • Familiarity with Python, to be able to read and modify scripts

Required Software

SSH Client to connect to Quest