Year: 2026
Type: preprint
Abstract: In Monte Carlo (MC) simulations particle moves require an O(N log N) particle mesh Ewald (PME) calculation for the treatment of long-range electrostatics, a calculation that is computationally demanding.We present the Precomputed Grid Potential (PGP) method, a mathematically exact recasting of the PME framework-trial cross-term that results in a reciprocal space complexity of O(1) for each insertion, without loss of accuracy.The fundamental principle of this approach is to exploit the fact that for a system with a stationary framework of atoms, the formally reciprocal electrostatic potential can be divided into static and dynamic contributions.After precomputing the static contribution of the stationary charges on a 3D grid employing a Bspline interpolation operator, particle insertions then only require O(1) interpolations of the stationary potential and not a computationally intensive Fourier transform.It is shown that the PGP and PME cross terms are algebraically the same and that the only source of numerical variation between them is the position-dependent mesh self-artifact included in discrete PME.This is eliminated by construction in PGP's cross-term formulation.For validation, we have compared PGP to PME for protein systems with up to 56K atoms, showing through production grand canonical Monte Carlo (GCMC) simulations that after the independent calibration of μex, all distribution measurements (e.g.mean occupancy, interaction energy, and acceptance ratio) were statistically indistinguishable from one another.Notably, computational times are cut by about two orders of magnitude, with no measurable loss of accuracy in the observables examined.
Source: ChemRxiv
Institution: University of Maryland, Baltimore
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
Funder: National Institutes of Health
Award: G9000096784
OpenAlex ID: W7160157808