HPC Curriculum and Associated Ressources in the Academic Context (1802.01110v1)

Abstract: Hardware support for high-performance computing (HPC) has so far been subject to significant advances. The pervasiveness of HPC systems, mainly made up with parallel computing units, makes it crucial to spread and vivify effective HPC curricula. Besides didactic considerations, it appears very important to implement HPC hardware infrastructures that will serves for practices, and also for scientific and industrial requests. The latter ensures a valuable connection with surrounding cutting-edge research activities in other topics ({\em life sciences, physics, data mining, applied mathematics, finance, quantitative economy, engineering sciences}, to name a few), and also with industrial entities and services providers from their requests related to HPC means and expertise. This aspect is very important as it makes an HPC Center becoming a social actor, while bringing real-life scenarios into the academic context. The current paper describes the major steps and objectives for a consistent HPC curriculum, with specific analyses of particular contexts; suggests how to technically set up operational HPC infrastructures; and discusses the connection with end-users, all these in both effective and prospective standpoints.