Day 4 - Advanced GPU programming optimisation - Paul Richmond

Description of session: Accelerators such as GPUs are prevalent both within personal computing as well as within high performance computing systems. Compared with traditional CPU architectures, performance is leveraged through a design which favours a high number of parallel compute cores at the expense of imposing significant software challenges. This optimisation session will provide strategies for improving the performance of GPU CUDA code. A hands-on lab session will provide an opportunity to try out the ideas presented.