Molecular Dynamics Simulation by Implementing Parallel Computing and Big Data Principles

Abstract

Several researchers are now able to run large-scale molecular dynamics simulations that can mimic the behaviour of a large number of atoms thanks to the emergence of big data and improvements in computing power. Large amounts of data are produced by these simulations, which can be processed, saved, and analysed using big data and cloud computing technologies. These algorithms can be applied to optimise the behaviour of quantum systems as well as forecast the characteristics of materials. We discuss the implementation of molecular dynamics simulations and aim to study the computational time for different number of atoms in simulation of a parallel computing model based on molecular dynamics. Molecular dynamics is an essential tool for understanding molecular reactions, and we aim to study the computational methods to make it faster and more accurate. However, such simulations may be computationally inefficient and memory-costly. The proposed solution is to parallelise the simulation algorithm, which could be a small step towards a larger solution. We propose a Charm++ based parallel algorithm that helps process large amount of big data and may see potential success in quantum computing.