DCAMM High Performance Computing, Open MP, and MPI

Title: High Performance Computing, Open MP and MPI
Teachers: Jens Honore Walther, Technical University of Denmark, and Ivo F. Sbalzarini, TU Dresden & MPI-CBG.
Time: January 16-20, 2017
Place: Technical University of Denmark, Lyngby, Denmark


Week 3: MPI

Lecturer: Ivo F. Sbalzarini
Exercise tutors: Pietro Incardona, Krzysztof Gonciarz
Exercise Solutions:
All solutions in one ZIP file


Monday:

Control structures, communication models, performance evaluation
Pipelining and vectorization, memory issues and pointers
Slides PDF, Handouts PDF
Exercise PDF


Tuesday:

Message passing, Fortran and MPI, blocking point-to-point, non-blocking, and collective communication.
Slides PDF, Handouts PDF
Exercise PDF


Wednesday:

Data types, groups, topologies, error handling, and parallel debugging
Slides PDF, Handouts PDF
Exercise PDF


Thursday

Domain decomposition, load balancing, hybrid particle-mesh methods
Slides PDF, Handouts PDF
Exercise PDF


Friday

Abstractions, parallel libraries
Slides PDF, Handouts PDF
Exercise PDF, Particle Positions DATA, Skeleton CODE


Links

Open MPI documentation
The MPI 2.2 Standard


Textbooks

MPI: The Complete Reference by Marc Snir, Steve Otto, Steven Huss-Lederman, David Walker, Jack Dongarra
Parallel Programming with MPI by Peter Pacheco