September 24, 2019 - by CSCS

The aim of this course is to introduce the participant to the HPX library for parallelism and concurrency and enable them to write applications using HPX. The library offers a C++ standards conforming API for task based programming, this is introduced and explained in detail with examples of DAG creation using Futures. Integrating futures with CUDA code and distributed applications. Mixing single node HPX with existing MPI based code and using executors for task placement.

The course is intended for confident C++ programmers - HPX makes use of templates in its API and the HPX internals use template metaprogramming methods to achieve function dispatch. Some knowledge of these techniques is an advantage. User may be interested in attending both the C++ training course (Mon-Wed) and the HPX course (Thu-Fri) back to back. The attendees must have a valid affiliation to an academic institution.

For the full agenda and registration, please visit the event page >