Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MM: Fachverband Metall- und Materialphysik

MM 9: Poster

MM 9.73: Poster

Montag, 17. März 2025, 18:30–20:30, P1

RuNNer 2.0: An Efficient and Modular Program for High-Dimensional Neural Network Potentials — •Alexander L. M. Knoll^1,2, Moritz R. Schäfer^1,2, K. Nikolas Lausch^1,2, Moritz Gubler³, Jonas A. Finkler³, Alea Miako Tokita^1,2, Gunnar Schmitz^1,2, Henry Wang^1,2, Richard Springborn^1,2, Marco Eckhoff⁴, and Jörg Behler^1,2 — ¹Theoretische Chemie II, Ruhr-Universität Bochum, Germany — ²Research Center Chemical Sciences and Sustainability, Research Alliance Ruhr, Germany — ³Department of Physics, Universität Basel, Switzerland — ⁴Laboratorium für Physikalische Chemie, ETH Zürich, Switzerland

Machine learning potentials (MLPs) have become an important tool for atomistic simulations in chemistry and materials science. As methods in this domain grow increasingly complex and mature, the creation of efficient and user-friendly libraries now receives a lot of attention. We introduce the second major release of RuNNer, an open-source, standalone software package designed for constructing and evaluating second-, third-, and fourth-generation high-dimensional neural network potentials (HDNNPs). RuNNer 2.0 integrates the entire workflow into a fully OpenMP- and MPI-parallel program: from generating atomistic descriptors, via training a specific machine learning model, to its application in molecular dynamics simulations.

Keywords: Software; Implementation; Machine Learning Potentials; High-Dimensional Neural Network Potentials