Register to this seminar by March 26, 2019.
Graphical Processing Units (GPUs) improve performance in video games or video editing. Their ability to simultaneously process many pixels ends itself well to parallel calculations.
GPUs are ideally suited for calculations that require many simple tasks on many data points simultaneously. Although the speed of a single core on a GPU is usually lower than for a CPU, the fact that a GPU has so many allows for a significant speed boost.
In this seminar, Jarno van der Kolk will show how to get started with programming for GPUs. Two frameworks exist to access the GPU, namely OpenCL and CUDA. We will touch on both, but the bulk of the seminar will be about CUDA.
We will also discuss how to run this code on the computational resources provided by Compute Canada, where you can have access to systems with up to 4 GPUs.
The code examples for this seminar are available on BitBucket.