For computational modeling and simulation, widely considered as the third paradigm of science, efficient use of computer hardware is vital. The same is also true for the newly resurgent field of machine learning and artificial intelligence. The Department of High-Performance Computing (HPC) thus sees as its mission to enable large and huge-scale computations in these contexts, through resource-efficient usage of modern computing platforms.
To achieve this ambitious goal, the HPC department will research several topics, including methodologies of parallel programming, hardware-compatible and/or inspired numerical strategies, software tools for user-friendly deployment and optimization of scientific code, plus real-world applications from various branches of computational science. The department will actively engage in multi-disciplinary collaborations across the departmental boundaries within Simula, as well as nationally and internationally.
The national research infrastructure eX3 (Experimental Infrastructure for Exploration and Exascale Computing), which is hosted at Simula, will be the primary hardware testbed for the HPC department. The various bleeding-edge processor architectures and interconnect solutions that are available on this extremely heterogeneous infrastructure will be carefully studied and tested for preparation of the upcoming Exascale Computing era and beyond. In particular, the machine-learning specific hardware within the eX3 platform is expected to pave the way for the forward-looking, multi-disciplinary research direction of scientific machine learning.