This course combines an overview of how to solve the design issues you will face when writing large-scale parallel programs, and a hands-on workshop on using the MPI parallel communication library to implement these solutions. It is suitable for those who have some knowledge of parallel programming, but would like to improve their skills and effectiveness. We strongly suggest you attend the previous day's 'Introduction to HPC & Parallel Programming Basics', unless you already have experience of using MPI and its core functions.
Topics covered will include:
- Domain Decomposition
- MPI Derived datatypes
- Latency hiding for improved performance
- Profiling parallel programs
- Discussing attendees' parallel programming applications.
The course runs as a full day event. The course is run consecutively with the introductory course 'Introduction to HPC & Parallel Programming Basics', running the previous day. It is intended that most course participants will register to attend the full 2-day program, as we believe this is the best way to get the most out of each of these courses.
Guest accounts will be provided for attendees that do not have an account on a VPAC Supercomputer, however, if you plan to get an account in the near future, please consider applying before the course and you will be able to use your own. Please mark on you account application forms that you will be attending this course and we will try and get your account created before hand.
This course is available to staff and students at VPAC Member Universities, however, booking in advance is essential. Please book your place by selecting a session time or contact firstname.lastname@example.org or phone +61 3 9925 4410.
This course will be held at 110 Victoria Street, Carlton (Ground Floor, Building 91, RMIT)