[Numerical Modeling 9] High-performance computing and parallel programming in Python

With multi-core processors available almost on every modern machine, as well as the availability of supercomputers with thousands to millions of processors, parallel programming has become essential to know if you want to improve the speed of execution of your computation dramatically. Actually, if you are interested to take advantage of the power of the modern processors, you need to break your problems into smaller blocks and distribute them among the available cores (like 8 cores of a multi-core Core i7 CPU). This is what parallel programming and high-performance computing (HPC) are all about. Although it may seem to be complicated, this video shows you that it is not.

Educational Materials:
In order to follow the videos, you need the educational materials, which are provided as a set of Jupyter Notebooks. You can find the materials at http://tuxriders.com/videos/numerical-computing/ and https://github.com/TuxRiders/numerical-computing-intro

Topics covered:
? Introducing to HPC and parallel programming
? Easy parallel computing using multiprocessing package
? Introducing IPython parallel
? Shared-memory and distributed-memory systems
? Message Passing Interface (MPI) for distributed-memory systems
? OpenMP for shared-memory systems
? Effect of problem size and overhead on the parallel performance

Lecturer: Mojtaba Barzegari https://mbarzegary.github.io/

To learn more about the goals of the TuxRiders project, please visit our website at http://tuxriders.com.

Chapters in this video!
################
00:00 - Intro
01:19 - Multiprocessing package
03:53 - IPython parallel
05:34 - Message passing interface (MPI)
08:54 - MPI examples
13:15 - OpenMP
16:05 - Effect of problem size on improved efficiency
18:07 - OpenCL