Numerical solution of scientific problems using GPUs and multicore architectures.
CUSO doctoral school is organizing a winter school FOR PHD STUDENTS dedicated to introduction of High Performance Computing and more specifically to GPU/multi-core computing.
This Winter School is a practical introduction to General-Purpose Graphics Processing Units (GPGPU) and multicore computing in the context of scientific research. The audience are researchers with introductory-level programming knowledge, but without necessary knowledge of high-performance computing. Modern and simplified approaches to parallel programming are presented in the Python, Julia, and C++ programming languages. A gentle introduction to deep learning will be given and concrete research applications will illustrate its complementarity with high-performance computing. Gonzalo Brito, our special guest, is a renowned researcher and developer of programming at NVIDIA. He actively participates in the development of new compiler technologies for GPU usage and participates in the improvement of the C++ standard is this domain. He is part of a very limited number of experts with experience in teaching the very recent parallel C++ standard.
The fees will be covered by CUSO (transport, hotel, and meals) excepted 50.- that will have to be paid by the participants (and hopefully reimbursed by their lab).