Parallel Implementation of Nonadditive Gaussian Process Potentials for Monte Carlo Simulations

A strategy is presented to implement Gaussian processpotentialsin molecular simulations through parallel programming. Attention isfocused on the three-body nonadditive energy, though all algorithmsextend straightforwardly to the additive energy. The method to distributepairs and triplets between processes is general to all potentials.Results are presented for a simulation box of argon, including fullbox and atom displacement calculations, which are relevant to MonteCarlo simulation. Data on speed-up are presented for up to 120 processesacross four nodes. A 4-fold speed-up is observed over five processes,extending to 20-fold over 40 processes and 30-fold over 120 processes.

Automated summary

This study presents a strategy to implement Gaussian process potentials in molecular simulations through parallel programming. The focus is on the three-body nonadditive energy, but all algorithms can be easily extended to the additive energy. The method to distribute pairs and triplets between processes is applicable to all potentials. Results are shown for a simulation box of argon, including full box and atom displacement calculations, which are relevant to Monte Carlo simulation. Data on speed-up are presented for up to 120 processes across four nodes, with a 4-fold speed-up observed over five processes, extending to 20-fold over 40 processes and 30-fold over 120 processes.