The course discusses programming techniques required to efficiently utilise parallel computing in a computational research project in science or engineering. The course will discuss shared memory and distributed memory parallelisation in a C, C++ and Fortran context. Widely utilised parts of the application interfaces of OpenMP and MPI will be introduced during the course. The course will discuss commonly encountered issues in parallel programming, such as data-races and dead-lock and show techniques required to avoid these issues.

Common programming tools will be introduced and demonstrated. This includes parallel profilers which are extremely helpful, when it comes to understanding performance problems in parallel and serial applications.

Prerequisites

Participants should be able to write simple programs in one or more of C, C++ or Fortran. 

Schedule

The course will start on the 4th of November 2024.  Teaching is expected to continue until February 2025.   Project presentation are planned for March 2025.   Students are expected to work on their projects during February/March 2025.

The teaching schedule is as follows:

• Mondays, 13.30 - 15.00, MH:228: Lecture
• Wednesdays: 13.30 - 15:00, MH:227: Programming exercises
• Thursdays: 10:15 - 12:00, MH:309B: Lecture

Teacher

Joachim Hein (Centre of Mathematical Sciences/LUNARC)

Registration

Please use the registration form to register for the course.

Registration will close on 18th October 2024.