Adaptively restrained molecular dynamics in LAMMPS

Journal of Computational Physics

2017

Self Cite

We consider Adaptively Restrained Langevin dynamics, in which the kinetic energy function vanishes for small velocities. Properly parameterized, this dynamics makes it possible to reduce the computational complexity of updating inter-particle forces, and to accelerate the computation of ergodic averages of molecular simulations. In this paper, we analyze the influence of the method parameters on the total achievable speed-up. In particular, we estimate both the algorithmic speed-up, resulting from incremental force updates, and the influence of the change of the dynamics on the asymptotic variance. This allows us to propose a practical strategy for the parametrization of the method. We validate these theoretical results by representative numerical experiments.

Section: Numerical Illustrationmentioning

confidence: 99%

Estimating the speed-up of adaptively restrained Langevin dynamics

Trstanova

Journal of Computational Physics

2017

Self Cite

“…To compute forces due to active interactions, we previously introduced active neighbor lists (ANLs), that is, lists of active neighbor particles at a given time‐step . The ANLs are constructed from full neighbor lists (FNLs), and the ANL of an active particle i can be described as

ANL (i) = {\forall j \in FNL (i) : j \in R \cup (j \in A \cap (j < i))}

.…”

Section: Algorithms For Ar Molecular Dynamicsmentioning

confidence: 99%

“…In contrast, in ARMD, only the positions of active particles are updated, and forces are updated based on the new positions of active particles, thereby eliminating the need to compute all forces. In our previous approach, active interactions were incrementally updated using an algorithm that involved two force update steps: First pass or subtraction step: This pass removes the force increments involving active particles based on the old positions, by computing the force increments due to active particles and subtracting them from the total force. Position update step: This involves updating the positions of active particles only, instead of updating the positions of all particles. Second pass or addition step: This step involves computing force increments based on the updated positions of active particles, which are then added to the forces obtained from the subtraction step. …”

Section: Algorithms For Ar Molecular Dynamicsmentioning

confidence: 99%

“…This incremental force update algorithm requires force increments to be calculated twice in a single integration step, limiting the usefulness of ARMD to cases where at least 50% of particles are restrained . Furthermore, if all particles are active, the proposed incremental force algorithm would be twice times slower than the normal MD algorithm.…”

Section: Algorithms For Ar Molecular Dynamicsmentioning

confidence: 99%

“…When two particles are restrained, the forces they apply on each other do not need to be updated from one time to the next. As a result, incremental force update algorithms, that is, force calculations that keep track of the total force applied on each particle and update forces involving active particles, may significantly speedup the simulation …”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Single‐pass incremental force updates for adaptively restrained molecular dynamics

Singh

2017

J Comput Chem

Self Cite

Adaptively restrained molecular dynamics (ARMD) allows users to perform more integration steps in wall-clock time by switching on and off positional degrees of freedoms. This article presents new, single-pass incremental force updates algorithms to efficiently simulate a system using ARMD. We assessed different algorithms for speedup measurements and implemented them in the LAMMPS MD package. We validated the single-pass incremental force update algorithm on four different benchmarks using diverse pair potentials. The proposed algorithm allows us to perform simulation of a system faster than traditional MD in both NVE and NVT ensembles. Moreover, ARMD using the new single-pass algorithm speeds up the convergence of observables in wall-clock time. © 2017 Wiley Periodicals, Inc.

Incremental update of electrostatic interactions in adaptively restrained particle simulations

Edorh

2018

J Comput Chem

The computation of long-range potentials is one of the demanding tasks in Molecular Dynamics. During the last decades, an inventive panoply of methods was developed to reduce the CPU time of this task. In this work, we propose a fast method dedicated to the computation of the electrostatic potential in adaptively restrained systems. We exploit the fact that, in such systems, only some particles are allowed to move at each timestep. We developed an incremental algorithm derived from a multigrid-based alternative to traditional Fourier-based methods. Our algorithm was implemented inside LAMMPS, a popular molecular dynamics simulation package. We evaluated the method on different systems. We showed that the new algorithm's computational complexity scales with the number of active particles in the simulated system, and is able to outperform the well-established Particle Particle Particle Mesh (P3M) for adaptively restrained simulations. © 2018 Wiley Periodicals, Inc.