GPU-Accelerated Implementation of Constant-pH Molecular Dynamics in NAMD
We introduce a GPU-based implementation of the hybrid nonequilibrium molecular dynamics-Monte Carlo constant-pH simulation method within the NAMD package. Benchmark results demonstrate that the new implementation achieves a substantial speedup over its CPU-based counterpart, while preserving the same level of accuracy. This performance gain expands the practical applicability of pH-controlled, all-atom simulations across a wide range of biomolecular systems. Although exhaustive enumeration of all possible protonation-state combinations remains a major challenge within conventional molecular dynamics frameworks, the ability to simulate at a defined pH─combined with GPU-acceleration─offers a powerful and efficient means for investigating dynamic protonation profiles of biomolecules.
Authors
- Sarah Moe (ORCID: https://orcid.org/0009-0009-1528-7079)
- Christophe J. Chipot (ORCID: https://orcid.org/0000-0002-9122-1698)
- Benoı̂t Roux (ORCID: https://orcid.org/0000-0002-5254-2712)
Institutions
- University of Illinois Urbana-Champaign (US)
- University of Chicago (US)
- Centre National pour la Recherche Scientifique et Technique (CNRST) (MA)
- Institute of Theoretical Physics (CN)
- Université de Lorraine (FR)
Publication Details
- Journal
- Journal of Chemical Information and Modeling
- Published
- 2025-12-09
- DOI
- https://doi.org/10.1021/acs.jcim.5c02162
- Primary Topic
- Protein Structure and Dynamics
- Type
- article
- Field-Weighted Citation Impact
- 0.00