Parallel N-body simulations of planetary systems: a direct approach

TL;DR

This paper presents a parallel N-body simulation of the solar system over a millennium, utilizing shared and distributed memory systems on a supercomputer to efficiently model gravitational interactions among celestial bodies.

Contribution

It introduces a direct parallel approach to N-body simulations of planetary systems, demonstrating scalability on high-performance computing architectures.

Findings

Successful simulation of solar system evolution over 1000 years

Effective parallelization on shared and distributed memory systems

Utilization of NASA data for initial conditions

Abstract

Direct gravitational simulations of n-body systems have a time complexity O(n^2), which gets computationally expensive as the number of bodies increases. Distributing this workload to multiple cores significantly speeds up the computation and is the fundamental principle behind parallel computing. This project simulates (evolves) our solar system for the next 1000 years (from 2015 to 3015) on the BlueCrystal supercomputer. The gravitational bodies (planets, moons, asteroids) were successfully simulated, and the initial states (mass, position and velocity vectors) of the solar system objects were obtained via NASA's JPL Horizons web interface. Two parallel computing domains are investigated: shared and distributed memory systems.