Computing Seminars/Talks: A quick introduction to parallel computing with MPI
15 March 2017 16:00 in CM301
Parallel programming for distributed memory machines (e.g. the hamilton cluster) is an indispensable skill for the modern applied mathematician. In this talk, I’ve tried to distill the “theoretical minimum” that you need to know to get started, using the most common interface: MPI. This is a library that works with either Fortran or C. The talk will be structured around a simple example of the one-dimensional heat equation. I’ve uploaded the example code (both Fortran 90 and C++ versions) to https://github.com/antyeates1983/mpi_seminar, in case you want to follow along.
Contact firstname.lastname@example.org for more information
This series of talks/seminars has the intention of sharing some practical computing skills in the department.
The talks are primarily aimed at PhD students in the early part of their PhDs, though they might also be of interest to students further through their PhDs and perhaps even others in the department. The idea is for older PhD students, and maybe other researchers, to share some tips and techniques that they’ve learnt from their own research in whichever software tools they use. It is hoped that this way we can share some useful tips and techniques and avoid new researchers having to spend a long time googling for ways to solve some common problems.
Supplementary files for these talks can be found here.