N-body Simulations of the Solar System with CPU-based Parallel Methods

TL;DR

This paper evaluates CPU-based parallel methods for simulating the Solar System's gravitational N-body problem, comparing computational efficiency and accuracy across different models and processor counts.

Contribution

It introduces and compares parallel approaches for N-body simulations, demonstrating their effectiveness in modeling the Solar System with acceptable accuracy.

Findings

Parallel methods improve simulation speed with increasing processors.

Second-order velocity Verlet provides acceptable orbital accuracy.

Performance varies with model size and processor number.

Abstract

The gravitational N-body simulation in the Solar system was performed using different parallel approaches with the comparisons in the computational times and speed-up values being carried out under different model sizes and the number of processors. The numerical integration used is a second-order velocity Verlet approach which gives the acceptable accuracy in the orbits of major bodies and asteroids with a time step size of 0.1 days.