In­tro­duc­ti­on to Ju­lia for High Per­for­mance Com­pu­ting

From December 5 to 7, 2023, a three-day course took place at the Paderborn Center for Parallel Computing (PC2), in which the Julia programming language was presented as a modern approach to high-performance computing. The event was aimed at developers who want to use HPC hardware efficiently without sacrificing the productivity of a dynamic high-level language. On the first day, the basics of Julia, including the compilation pipeline and the multiple dispatch paradigm, were presented. After that, the course focused on the HPC aspects of sequential and parallel programming in Julia. Among others, the following topics have been covered: type and memory optimizations, SIMD, multithreading, distributed computing, and GPU acceleration. In addition, the course offered Julia and HPC enthusiasts the opportunity to network and share their experiences.

The course proved very popular and attracted more than 100 registrations, of which 50 were ultimately accepted. It was the first fully hybrid training event in the brand new training room in Building X of PC2 with online participants from all over Germany and on-site participants from Aachen, Dortmund, Münster and Wuppertal.

The course was a joint event of the Paderborn Center for Parallel Computing, the NHR Alliance and the HPC.NRW competence network.