Assignment 4: Parallelization and heterogeneous platforms

This assignment further extends the previous one. Last time we were looking for single core performance, this time we try to distribute the algorithm over multiple cores to gain even more performance. This assignment is also planned to be in two parts, but currently only the first part is finished.

Assignment 4a: Parallelization using OpenMP

We apply the OpenMP programming model which was introduced during the lectures and demonstrate its capabilities for parallelizing an algorithm over multiple processors. This time we focus on the traditional OpenMP parallelization for homogeneous multi-core systems.

Assignment 4b: Utilizing heterogeneous hardware platforms

Under construction Extending on the previous part we now introduce accelerator offloading and try to map parts of the image processing algorithm onto a GPU.