Nature Unknown label
Credit hour 3
Total number of hours 20
Number of hours for lectures 20

Content

High Performance Computing (HPC) has moved from a selective and expensive
endeavor to a cost-effective technology within reach of virtually every budget with
the massive arrival of many cores processors (APU, GPU…). The purpose of this course
is to introduce Scientific computing, parallel programming architectures and models
deployed on supercomputers, computing clusters and grids. Hybrid computing is also
presented with modern accelerators, including General Purpose Graphical Processing
Units, manycore and also disruptive technologies like photonic and quantum
computing. This course presents the history, the theoretical laws, the modern
concepts and the future of High Performance Computing. It gives a particular focus on
stochastic distributed computing and parallel random numbers and on numerical
reproducibility which provide a robust and resilient approach to address the increase
in soft errors met on our roadmap towards Exascale performances (2022).