High Performance & Parallel GPU Computing: Summer School 1285 views

Institute of Computational Science at USI and the Swiss National Supercomputing Centre (CSCS): applications are invited for CSCS-USI Summer School 2018. The summer school will focus on the effective exploitation of hybrid High Performance Computing (HPC) systems, equipped with powerful CPUs, state-ofthe-art GPUs, and fast and scalable networks.

About the program:

• Experts from ETH Zurich, Università della Svizzera italiana, and NVIDIA will present CUDA and OpenACC for GPU programming and MPI for inter-GPU/CPU communication.
• They will also introduce scientific libraries and machine learning frameworks, optimized for the GPU and for parallel applications.
• Extensive practical and exercise lab sessions will help to clarify and consolidate the theoretical material.
• The Summer School addresses undergraduate students, Ph.D. students, Postdocs and other researchers.
• Selected students will be able to earn three ECT credit points for this course from Università della Svizzera italiana (subject to exam).

Duration: July 15 – July 27, 2018

Location: Steger Center, Riva San Vitale, Ticino, Switzerland

Study areas:

• GPU architecture
• GPU programming with CUDA and OpenACC
• Message passing programming model, the MPI standard and the GPUDirect library
• Performance optimization
• Scientific libraries for parallel and GPU programming
• GPU optimized machine learning frameworks