Efficient Scheme for Active Particle Selection in N-body Simulations

TL;DR

The paper introduces an efficient active particle selection method for N-body simulations that significantly reduces computational complexity by focusing on the average active particles per step, saving time especially when this number is small.

Contribution

It presents a novel method that decreases the computational complexity of active particle selection in N-body simulations from O(N·N_{step}) to O(average_active_particles·N_{step}), improving efficiency.

Findings

Reduces active particle selection time in simulations.

Effective when the average active particles per step are small.

Significantly improves simulation efficiency in low activity scenarios.

Abstract

We propose an efficient method for active particle selection, working with Hermite Individual Time Steps (HITS) scheme in direct N-body simulation code $\varphi$GRAPE. For a simulation with $N$ particles, this method can reduce the computation complexity of active particle selection, from $O(N\cdot N_{step})$ to $O(\overline{N_{act}}\cdot N_{step})$, where $\overline{N_{act}}$ is the average active particle number in every time step which is much smaller than $N$ and $N_{step}$ is the total time steps integrated during the simulation. Thus can save a lot of time spent on active particle selection part, especially in the case of low $\overline{N_{act}}$.