Speaker: Dr. Carsten Bauer, Paderborn Center for Parallel Computing (PC2)

Slides: https://hpc.fau.de/files/2024/05/bridging_science_hpc.pdf

Abstract:

Julia is a modern programming language that is particularly well suited for numerical computing. Its rich scientific ecosystem and dynamic features (e.g., interactivity) make it highly appealing to specialized domain scientists. At the same time, thanks to its LLVM-based “just ahead of time” compiler, Julia can often offer state-of-the art performance on modern CPUs and GPUs. The combination of these two facets makes it a promising option for HPC with intriguing technical and social potential. In practice, however, synergy effects are sometimes coupled with shortcomings, some of which can probably be overcome with increasing maturity, while others appear to be more fundamental.

In this HPC Café talk, I will share my experiences with Julia as a physicist, as an HPC consultant within the NHR, and as a long-time member of the Julia community. I will discuss the present state of the language and its potential from a HPC perspective while also shedding light on several noteworthy international Julia HPC projects.

Material from past events is available at: https://hpc.fau.de/teaching/hpc-cafe/