NIC Symposium 2022

NIC Symposium 2022 Proceedings

Proceedings NIC Symposium 2022, Jülich, Germany
29 Sept 2022 - 30 Sept 2021, pdf

On September 29 and 30, 2022, computational scientists will convene in Jülich for the 11th NIC symposium to present their exciting research – illustrating the diverse range of modern, computational science at the John von Neumann Institute for Computing (NIC). For many of us, the previous NIC symposium in February 2020 was the last face-to-face event before the COVID-19 pandemic struck, and we have shifted the present symposium to the fall in order to allow for a lively in-person exchange. It has been a tradition that the biannual symposia are accompanied by proceedings that address both, computational scientists and practitioners as well as the general public, interested in the advancement of computational science and its applications in diverse, contemporary research fields. We are delighted that also this year it was possible to showcase the breadth of high-performance computing research, with contributions from elementary particle physics, astrophysics, and statistical physics of hard and soft condensed matter, computational chemistry and life sciences, as well as material science, fluid-dynamics engineering, and climate research. Since its foundation 24 years ago, the NIC continuously provides the scientific community with essential high-performance computing resources, training, and technical support on the highest performance level. Within the Gauss Centre for Supercomputing (GCS) the Jülich Supercomputing Centre (JSC) provides a modular supercomputer, comprised of a versatile, easy-to-use, CPU-based cluster module that has been installed 2018 and a GPU-based booster module. The latter has been deployed in 2020 and features more than 3700 NVIDIA A100 GPUs. By virtue of the efficient, dedicated and much appreciated training and user support of the JSC, this new architecture has been swiftly adopted across the different scientific communities. This has been an important step in moving forward the frontiers of established simulation techniques e.g., in elementary particle physics or extremely large, particle-based simulations in materials physics and engineering and has also enabled new applications using big scientific data analysis and machine-learning strategies. For instance, in the MLPerf Training HPC competition in November 2021 researchers of the JSC and the Steinbuch Centre for Computing at the Karlsruhe Institute of Technology (Helmholtz AI) were able to perform the fastest ever and most computationally intensive AI calculations in Europe.

Last Modified: 30.09.2022