A method and system for calculating the interaction force between particles

Abstract

The invention discloses a method and a system for calculating the interaction force between particles, and relates to the technical field of molecular dynamics simulation systems. The charge distribution and the charged amount of the particle are used to obtain the Coulomb long-range force of the particle; wherein, the Fourier space corresponds to the function for calculating the Coulomb long-range force. Sampling methods include importance sampling, where fractional Fourier terms are randomly drawn on the order of hundreds or thousands. The method also includes obtaining the Coulomb short-range force according to Coulomb's law; obtaining the Leonard Jones force and bonding interaction force of the particle according to other parameters; finally obtaining the interaction force. The method can be used in molecular dynamics simulation to reduce computational time complexity and data traffic, thereby improving computational performance and parallel efficiency, and reducing communication consumption.

Description

technical field [0001] The invention relates to the technical field of molecular dynamics simulation, in particular to a method and system for calculating the interaction force between particles. Background technique [0002] Molecular dynamics simulation is an important computer simulation method to describe the evolution of micro- and nano-scale systems. It is widely used in the fields of biomolecules, drug design, and energy materials. It obtains the velocities and positions of all particles by calculating the interaction force between particles and Newton's equation of motion, and obtains a series of thermodynamic and dynamic properties from these trajectories, such as density distribution, binding energy, diffusion coefficient, mobility, etc. Traditional molecular dynamics force fields decompose the potential energy function into short-range bond energy interactions, van der Waals interactions, and long-range Coulomb interactions, and sequentially calculate their contri...

AI Technical Summary

Problems solved by technology

When implementing these methods, complex inter-process communication processing is required, which seriously reduces the parallel efficiency of these methods.

Another disadvantage of the FMM approach is the excessive processing time required to build complex data structures

The classic Ewald summation method needs to transmit the amount of data for each calculation is O(N 3 / 2 ), which means that as the number of particles increases, the communication time grows rapidly, leading to a decrease in the parallelism efficiency of these methods

As an FFT-based method, the PPPM method requires extensive and intensive communication between the CPU and the CPU, and the complexity is O(NlogN), which cannot achieve the optimal linear complexity.

Images