Oral examination (written exam if there are a large number of participants), successful exercises can result in a grade bonus

The lecture introduces the necessary concepts of parallel processing, and gives an overview of the currently available hardware. Furthermore, basic parallel algorithms, e.g. Map, Reduce, Prefix Sum, Branching, but also parallel applications like FFT, particle systems and simulations etc. are covered. In order to develop efficient parallel solutions for new problems, appropriate approaches and complexity analyses will be taught.

A current trend of all chip manufacturers is to integrate more and more computing units on one chip, e.g. with several hundred processors on one graphics card. In order to use these architectures efficiently, suitable algorithms must be chosen and the problems optimised in terms of memory bandwidth. (1) The aim of the lecture is to enable the students to analyse a given problem with regard to the possible increase in efficiency through parallelisation. (2) They are able to develop suitable algorithms to work out a massively parallel implementation as fast as possible. (3) They are able to optimise their programs in terms of memory bandwidth, GPU utilisation and registers by profiling.

Hubert Nguyen: GPU Gems 3, Addison Wesley; T. Mattson, B. Sanders, B.
Massingill: Patterns for Parallel Programming, Addison Wesley ; gpgpu.org
- General-Purpose Computation Using Graphics Hardware; NVIDIA CUDA
page; NVIDIA CUDA Programming Guide ; Vorlesungsfolien werden bereitgestellt