Skip to 0 minutes and 10 secondsHi, welcome to the final week of Python in high-performance computing course! In the previous weeks we have been looking on how to optimize your Python program using a single CPU core. This week we will look into going beyond a single core using parallel programming in Python. After optimizing the single core performance of your Python program, as we have been doing in the past weeks to gain further speed ups you can look into using multiple CPUs in parallel. Another reason for using multiple CPUs in parallel is if your data is large enough that it won't fit in the memory of a single CPU, by using multiple CPUs you get more memory for you to play with.

Skip to 1 minute and 3 secondsIn high-performance computing the de facto standard for using multiple CPUs is something called message passing interface, or MPI for short. MPI is used to exchange information by sending and receiving messages. In this week we will look into using MPI to do parallel computing using a package called mpi4py.

Welcome to week 4

In order to utilize multiple CPU cores, parallel processing is required.

During this week we discuss how parallel programming within the message passing paradigm can be done with Python and mpi4py package.

Share this video:

This video is from the free online course:

Python in High Performance Computing

Partnership for Advanced Computing in Europe (PRACE)