The Graphics Processing Unit or GPU is nowadays a mainstream component in Scientific Computing. For relatively little money one can have supercomputer performance. However, some extra work has to be done to make an ordinary sequential program suitable for use on the GPU. One of the most important tools for using GPUs is currently “CUDA” (Compute Unified Device Architecture). This is basically an extension to the C programming language, which can be used to program the GPU in an easy way. Another tool is OpenCL, but as this is less mature than CUDA and performs less on especially the newest GPUs, the focus will be currently on CUDA. During this lecture we will explain the basic principles of GPU programming. For a more in depth course, please see this link.

Cornelis Vuik is a full professor in Numerical Analysis at the Delft University of Technology. Furthermore he is Scientific Director of 4TU.AMI Applied Mathematics Institute and Director of TU Delft Institute for Computational Science and Engineering (DCSE). You can read more about him on his website.