Result: An automated parallel program of Clean Numerical Simulation for chaotic systems governed by ODEs.

Title:
An automated parallel program of Clean Numerical Simulation for chaotic systems governed by ODEs.
Authors:
Zhang, Bo1 (AUTHOR), Liao, Shijun1,2 (AUTHOR) sjliao@sjtu.edu.cn
Source:
Computer Physics Communications. Dec2025, Vol. 317, pN.PAG-N.PAG. 1p.
Subject Terms:
*CHAOS theory, *ORDINARY differential equations, *APPROXIMATION error, *AUTOMATION software, *PARALLEL programming, *COMPUTER software, *HIGH performance computing, *COMPUTER simulation
Database:
Academic Search Index

Further Information

Due to the butterfly-effect, numerical noise in chaotic systems grows exponentially, presenting a significant challenge. This issue can be mitigated through the use of Clean Numerical Simulation (CNS) proposed by Liao in 2009, which can effectively reduce numerical noise to a desired (say, arbitrarily low) level in an interval of time that is long enough for statistics. In this paper, we propose the CNSPy, a novel, highly efficient, self-adaptive CNS implementation to obtain the convergent (i.e. reproducible) numerical simulation of chaotic systems governed by a set of ordinary differential equations (ODEs). This software automates the CNS computational workflow by automatically converting Python-defined ODEs into a parallelized C code, eliminating the need for error-prone manual derivation and coding while ensuring high efficiency in high-performance computing (HPC) environments. The code is free and available at https://github.com/sjtu-liao/cnspy. [ABSTRACT FROM AUTHOR]