Introduction to Compute Unified Device Architecture (CUDA) for General Purpose GPU programming
by
Kyriakos Hadjiyiannakou(University of Cyprus)
→
Europe/Berlin
Zoom
Zoom
Description
This is an introduction to the Compute Unified Device Architecture (CUDA) introduced from Nvidia for a general purpose programming on GPUs. We introduce the concept of heterogeneous computing and
discuss the meaning of CUDA threads and blocks paying attention to the GPU architecture. We provide examples of vector additions and dot products in CUDA C/C++ to allow us better understanding of
CUDA. Additionally, we discuss more advanced features such as CUDA streams to overlap computation with communication.
