Workshops and courses
Programming on GPUs @MdS
Date: 7 - 9 December 2015
This course presents the basic of parallel programming on GPU architectures. Its main focus is on the CUDA language, but it will provide a brief overview of the main tools available on those platforms.Content
Introduction to parallel computing for GPU architectures.
CUDA programming model: thread grid hierarchy, memory model, CUDA-C language, massively multi-thread architecture, execution model
Basic GPU algorithms : sorting and reductions
Advanced aspects: performance optimization, profiling, productivity and scientific libraries (e.g. Thrust, Kokkos).
Half of the course will be made of hands-on sessions.Learning outcomes
Understanding of the main issues in developing GPU applications. Ability to choose the most suitable tool for a given application and environment. Ability to develop a working prototype for a target application.Prerequisites
Basic knowledge of C programming in a Unix environment.
Event types:
- Workshops and courses
Activity log
