Introduction to parallel programming with MPI and OpenMP
Lecturer: Dr. Hinnerk Stüben
Location: Online course
Date: 12.-15.04.2021, 10:00-16:30
Abstract
The course is aimed at students and scientists who are interested in parallelizing numerical simulation software. It covers the standard programming methods for that purpose: MPI for programming distributed memory machines (e.g. HPC clusters) and OpenMP for programming shared memory nodes. Lectures are supplemented by programming exercises.
Prerequisites
- programming experience in C/C++ or Fortran
- familiarity with Linux
Contents
- principles of parallel computing
- parallel computer architectures
- parallel programming models
- efficiency of parallel programs
- data dependence analysis
- programming with MPI (Message Passing Interface)
- messages
- point-to-point communication
- collective communication
- data types
- reduction operations
- one-sided communication
- parallel I/O
- programming with OpenMP
- threads
- synchronization
- parallelization of loops
- hybrid programming
- combining MPI with OpenMP
Registration
- by e-mail to the [organizer]