Faster Molecular Dynamics with Neural Network Potentials via Distilled Multiple Time-Stepping and Nonconservative...

Following our previous work [Cattin, C. J. Phys. Chem. Lett. 2026, 17(5), 1288-1295], we propose the DMTS-NC approach, a distilled multi-time-step (DMTS) strategy using nonconservative (NC) forces, to further accelerate atomistic molecular dynamics (MD) simulations using foundation neural network models such as FeNNix-Bio1. Therein, a dual-level reversible reference system propagator algorithm (RESPA) formalism couples a target accurate conservative potential to a simplified distilled representation optimized for the production of nonconservative forces. Despite being nonconservative, the distilled architecture is designed to enforce key physical priors, such as equivariance under rotation and cancellation of atomic force components. These choices facilitate the distillation process and therefore improve drastically the robustness of simulation, significantly limiting abnormal discrepancies between the two models, thus achieving excellent agreement with the forces data. Overall, the DMTS-NC scheme is found to be more stable and efficient than its conservative counterpart with additional speedups reaching 15-30% over DMTS. Requiring no fine-tuning steps, it is easier to implement and can be pushed to the limit of the system's physical resonances to maintain accuracy while providing maximum efficiency. We obtain additional speedup by combining hydrogen mass repartitioning (HMR) and High Hydrogen Friction (HHF) to further extend the largest time step up to 10 fs of our schemes while conserving stability and accuracy. As for DMTS, DMTS-NC is applicable to any neural network potential (NNPs) and can be applied to approaches that are computationally heavier than FeNNix-Bio1. We show a proof of principle applying the approach to the distillation of MACE-OFF23 with consequent speedups ranging from 3.66 to 5.64 compared to a single time step.