Parallel computing

Welcome to the Coding for Reproducible Research parallel computing course. By the end of this course you will:

• be able to write software which can run on multiple-processes using MPI
• be able to write code that can run on multiple threads
• be able to identify how a software problem can be broken down and parallelised
• have a better understanding of distributed and shared-memory parallelism and how they should be employed

Course content

This course gives a high-level overview of some parallel computing paradigms, along with practical examples that will provide an entry point for participants to get started writing code that can make use of parallel architectures. The first practical portion of the course involves using MPI and the mpi4py Python package to get an overview of distributed memory parallelism. The second practical portion involves using what we learned about MPI to solve a problem in parallel using MPI.

Session 1

MPI installation instructions

1. MPI Hello world
2. Simple communications with MPI
3. Collective communication with MPI
4. Solving a problem in parallel

We ask that participants in this course please familiarise themselves with the code of conduct.