High-performance training and inference for deep equivariant interatomic potentials
Digital Discovery · 2026 · 5 (4) · 1558-1567
DOI: 10.1039/d5dd00423c ↗Kurz gefasst
Fachartikel in der Zeitschrift Digital Discovery, 2026. Verfasst von Chuin Wei Tan und 13 weiteren Autor:innen. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 71 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0). Gefördert von: National Science Foundation, Harvard University, U.S. Navy.
Abstract (Originalsprache)
The NequIP framework is redesigned for scalable distributed training and PyTorch 2.0 compilation. AOT Inductor inference and optimized Allegro kernels accelerate molecular dynamics by factors of 5–18 on practical system sizes.
Förderung
- National Science Foundation
- Harvard University
- U.S. Navy
- Office of Naval Research
- U.S. Department of Energy
- Basic Energy Sciences
- Robert Bosch
- Center for the Environment, Harvard University
Zitieren
APA
Tan, C. W., Descoteaux, M. L., Kotak, M., Nascimento, G. D. M., Kavanagh, S. R., Zichi, L., Wang, M., Saluja, A., Hu, Y. R., Smidt, T., Johansson, A., Witt, W. C., Kozinsky, B., & Musaelian, A. (2026). High-performance training and inference for deep equivariant interatomic potentials. Digital Discovery, 5(4), 1558-1567. https://doi.org/10.1039/d5dd00423c
Vancouver
Tan CW, Descoteaux ML, Kotak M, Nascimento GDM, Kavanagh SR, Zichi L, et al. High-performance training and inference for deep equivariant interatomic potentials. Digital Discovery. 2026;5(4):1558-1567. doi:10.1039/d5dd00423c
BibTeX
@article{Tan2026,
author = {Tan, Chuin Wei and Descoteaux, Marc L. and Kotak, Mit and Nascimento, Gabriel de Miranda and Kavanagh, Seán R. and Zichi, Laura and Wang, Menghang and Saluja, Aadit and Hu, Yizhong R. and Smidt, Tess and Johansson, Anders and Witt, William C. and Kozinsky, Boris and Musaelian, Albert},
title = {High-performance training and inference for deep equivariant interatomic potentials},
journal = {Digital Discovery},
year = {2026},
volume = {5},
number = {4},
pages = {1558--1567},
doi = {10.1039/d5dd00423c}
}
EUVIMED