RuNNer 2.0: A Software Suite for High-Dimensional Neural Network Potentials

Abstract

We present RuNNer 2.0, the "Ruhr University Neural Network energy representation", a highly optimized software suite for training and evaluating high-dimensional neural network potentials (HDNNPs) of the second, third, and fourth generation. Long-range electrostatics and charge equilibration (QEq) for the description of non-local charge transfer in fourth-generation (4G) HDNNPs are accelerated by quasi-linear-scaling plane-wave methods, reducing QEq computational complexity from 𝒪(N³) to 𝒪(Nlog² N) such that linear or quasi-linear scaling is achieved across all HDNNP generations. An optimized memory management strategy eliminates the training overhead traditionally associated with long-range interactions, allowing 4G-HDNNPs to be trained with the same efficiency as their local counterparts. Developed in modern Fortran (2003/2008 standards), combined with a hybrid MPI/OpenMP parallelization scheme, RuNNer 2.0 has been designed to run efficiently in any CPU environment, from cost-effective local workstations to massive HPC clusters. Its modular library architecture facilitates straightforward binding to external simulation software; native interfaces to LAMMPS and the Atomic Simulation Environment (ASE) provide full access to all its features, including built-in committee-based uncertainty quantification. The high efficiency and scalability of the RuNNer 2.0 ecosystem are demonstrated through detailed benchmarks.

Training of Neural Network Machine Learning Potentials for Activation Barrier Calculations in Predicting Growth Modes of Deposited Metals
Karl Rönnby, Cara-Lena Nies, Savannah May Timony +2
ChemRxiv6 Nov 2026PreprintLicence not statedGeneral chemistryChemRxiv
Why Multi-Layer Message Passing Works: Completeness Theory for Graph Neural Network Interatomic Potentials
Pingbing Ming, Han Wang
arXiv2 Sep 2026PreprintCC BY-NC-SA 4.0Full textPhysical chemistry Mathematical physicsarXiv
Learning Memory-Dependent Neural Network Correlation Potentials for Accurate Electron Dynamics
Harish S. Bhat, Christine M. Isborn
arXiv3 Sep 2026PreprintarXiv licenceFull textPhysical chemistry Optimisation & controlarXiv
Leveraging neural network interatomic potentials for a foundation model of chemistry
So Yeon Kim, Yang Jeong Park, Ju Li
npj Computational Materials17 Jun 2026Peer-reviewedCC BYFull text12 citedMaterials & polymersOpenAlex
AdaptNTK: Adaptive Uncertainty Quantification and Active Learning for Neural Network Potentials
Prajwal Ananth, Shuwen Yue
arXiv1 Sep 2026PreprintarXiv licenceFull textPhysical chemistryarXiv

Cite

BibTeX RIS

BibTeX

@misc{knoll2026runner,
  title = {{RuNNer 2.0: A Software Suite for High-Dimensional Neural Network Potentials}},
  author = {Alexander L. M. Knoll and Moritz R. Schäfer and K. Nikolas Lausch and Moritz Gubler and Henry Wang and Richard Springborn and Redouan El Haouari and Alea Miako Liebetrau and Jonas A. Finkler and Emir Kocer and Marco Eckhoff and Stefan Goedecker and Gunnar Schmitz and Jörg Behler},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2607.17978},
  eprint = {2607.17978},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2607.17978},
  note = {CC BY-SA 4.0},
}

RIS

TY  - UNPB
TI  - RuNNer 2.0: A Software Suite for High-Dimensional Neural Network Potentials
AU  - Alexander L. M. Knoll
AU  - Moritz R. Schäfer
AU  - K. Nikolas Lausch
AU  - Moritz Gubler
AU  - Henry Wang
AU  - Richard Springborn
AU  - Redouan El Haouari
AU  - Alea Miako Liebetrau
AU  - Jonas A. Finkler
AU  - Emir Kocer
AU  - Marco Eckhoff
AU  - Stefan Goedecker
AU  - Gunnar Schmitz
AU  - Jörg Behler
PY  - 2026
DA  - 2026/09/28
PB  - arXiv
DO  - 10.48550/arxiv.2607.17978
UR  - https://arxiv.org/abs/2607.17978
AB  - We present RuNNer 2.0, the "Ruhr University Neural Network energy representation", a highly optimized software suite for training and evaluating high-dimensional neural network potentials (HDNNPs) of the second, third, and fourth generation. Long-range electrostatics and charge equilibration (QEq) for the description of non-local charge transfer in fourth-generation (4G) HDNNPs are accelerated by quasi-linear-scaling plane-wave methods, reducing QEq computational complexity from 𝒪(N^3) to 𝒪(Nlog^2 N) such that linear or quasi-linear scaling is achieved across all HDNNP generations. An optimized memory management strategy eliminates the training overhead traditionally associated with long-range interactions, allowing 4G-HDNNPs to be trained with the same efficiency as their local counterparts. Developed in modern Fortran (2003/2008 standards), combined with a hybrid MPI/OpenMP parallelization scheme, RuNNer 2.0 has been designed to run efficiently in any CPU environment, from cost-effective local workstations to massive HPC clusters. Its modular library architecture facilitates straightforward binding to external simulation software; native interfaces to LAMMPS and the Atomic Simulation Environment (ASE) provide full access to all its features, including built-in committee-based uncertainty quantification. The high efficiency and scalability of the RuNNer 2.0 ecosystem are demonstrated through detailed benchmarks.
ER  -

Abstract

Related

Cite