This course is an introduction to the MPI paradigm to parallelize program running on distributed memory systems.
A short reminder of HPC architecture will be presented. The first part of the training will be dedicated to the concept of parallelization. Starting from a fully sequential program, some aspects commonly encountered during the parallelization of a program will be presented (such as finding independent areas within the program, code modifications to be introduced, how to manage I/O ...). These different aspects will be illustrated through several examples. The Message Passing Interface Library will be presented. We will describe the main MPI calls to make the beginner able to parallelize a Fortran 90 code using MPI library.
Dedicated to beginners, this training will be structured as a sequence of lectures and hands-on tutorial exercises. The exercises will be given in Fortran 90.  Learning outcomes
Participants will become familiar with the concept of parallelization and be able to parallelize a program in Fortran using basic MPI calls.Prerequisites
 Linux and Fortran 90 basics

https://events.prace-ri.eu/event/120/