Efficient implementation of periodic boundary conditions in Monte Carlo simulation

Area of Science:

• Computational chemistry
• Molecular dynamics

Background:

• Calculating shortest distances between particles is a bottleneck in Monte Carlo (MC) molecular simulations.
• Periodic boundary conditions (PBCs) are essential for simulating bulk systems but computationally intensive.

Purpose of the Study:

• To develop an optimized method for handling periodic boundary conditions in MC simulations.
• To improve the computational efficiency of molecular simulations.

Main Methods:

• Implemented signed-integer coordinate storage within a scaled box for PBCs.
• Reduced conditional statements in the PBC algorithm for cubic and rectangular boxes.
• Tested the optimized method using NVT Monte Carlo simulations.

Main Results:

• The signed-integer storage method successfully bypassed multiple conditional statements.
• Significant performance increase observed in NVT MC simulations.
• Validated the method with simulations of liquid krypton and water.

Conclusions:

• The proposed signed-integer coordinate storage is an effective technique for accelerating molecular simulations.
• This optimization offers a practical improvement for computationally demanding tasks in computational chemistry.