Chaoticus: a parallel approach to the computation of chaos indicators

TL;DR

Chaoticus is a GPU-accelerated Python package that efficiently computes chaos indicators for dynamical systems, enabling large-scale analysis of complex Hamiltonian dynamics.

Contribution

It introduces a novel GPU-based parallel implementation for computing multiple chaos indicators, significantly speeding up analysis compared to traditional CPU methods.

Findings

Reduces computation time by several orders of magnitude.

Enables generation of extensive datasets for complex dynamics analysis.

Supports multiple chaos indicators including SALI, GALI, and Lyapunov exponents.

Abstract

In this paper we present Chaoticus, a Python-based package for the GPU-accelerated integration of ODE systems and the computation of chaos indicators, including SALI, GALI, Lagrangian Descriptors based indicators and the Lyapunov exponent spectrum. By leveraging GPU parallelization, our package significantly reduces the computation times by several orders of magnitude compared to CPU-based approaches. This significant reduction in computing time facilitates the generation of extensive datasets, crucial for the in-depth analysis of complex dynamics in Hamiltonian systems.