Message Passing is presently the most widely deployed programming model in massively parallel high performance computing. Message passing is suitable for programming a wide range of current computer architectures, ranging from multi-core desktop equipment to the fastest HPC systems in the world, offering several hundred thousand processing elements.
This online course will teach you how to write parallel programs using MPI. The course will be delivered by PDC staff.
The course is at beginners level and assumes no prior experience in parallel computing. The concepts behind message passing and distributed memory computing will be introduced and the syntax of the key MPI calls will be explained. The course will include point-to-point communications, non-blocking communication and collective communications calls.
Presentations and type-along sessions will be interspersed with practical exercises where participants get experience in writing MPI programs and gain an understanding different parallelisation strategies. Code examples and exercises will be available in C, Fortran and Python. At the end of the course, participants should be able to write their own MPI programs at a basic level. The teaching language for the course will be English.
Participants can choose between setting up an MPI environment on their own computers or using PDC clusters if they already have a PDC account. Installation instructions will be sent out to registered participants.
Participants should be able to write programs in either C, Fortran or Python.
08:30 0. (Optional) Software installation help
09:00 1. Introduction to Parallel Computing
09:45 2. Serial and Parallel Regions
10:05 3. MPI_Send and MPI_Recv
10:30 4. Coffee Break
10:50 5. Parallel Paradigms and Parallel Algorithms
11:10 6. Non-blocking Communication
11:40 7. Collective Operations
12:10 8. Lunch Break
13:10 9. Serial to Parallel
14:10 10. Tips and Best Practices
14:30 11. Discussion and independent work
The course material can be found at https://pdc-support.github.io/introduction-to-mpi.
For more information, contact PDC Support: firstname.lastname@example.org.