Implementation of Parallel Algorithms for Liquid Metal Solidification Molecular Dynamics Based on Big Data

With the advancement of science and technology and the popularization of the Internet, large amounts of data and information will be generated from people's daily work and life to the state of molecular motion trajectories. How to efficiently mine the hidden information of these massive data information is the key to scientific and social development, which requires big data analysis technology. The solidification process of metals is a very complex process, including both macroscopic and microscopic processes. Molecular dynamics can simulate the process of liquid metal solidification from a microscopic scale. In molecular dynamics simulation, the more particles the simulation system contains, the more accurate the data obtained. However, the molecular dynamics serial program runs slowly, and the number of simulated atoms is limited, which makes the simulation results far from the real situation. The rise of parallel computing provides conditions for large-scale molecular dynamics parallelization and promotes the rapid development of molecular dynamics parallelization. The purpose of this paper is to study the dynamic parallel algorithm of liquid metal solidification molecules based on big data. In this paper, through in-depth research on big data, analysis of the current research results of parallel computing of liquid metal solidification molecules, and combined with the current research status of liquid metal solidification molecules in my country, to discuss the research of liquid metal solidification molecular dynamics parallel algorithm under big data. Research shows that the spatial domain decomposition method is more general, and its communication cost is relatively small. It reflects that it can obtain relatively high parallel performance and computing efficiency in local communication, especially in large-scale simulation systems with a large number of atoms.