Javascript must be enabled for the correct page display

PanPelagos-Core — Developing a Fast and Flexible Lagrangian Fluid Transport Simulator for Advanced Physics Applications

Hougee, Djairo and Opat, Martin and Fragiacomo, Fabio (2025) PanPelagos-Core — Developing a Fast and Flexible Lagrangian Fluid Transport Simulator for Advanced Physics Applications. Master's Internship Report, Computing Science.

[img]
Preview
Text
mCS2025Hougee-Fragiocomo-Opat.pdf

Download (930kB) | Preview
[img] Text
Toestemming.pdf
Restricted to Registered users only

Download (157kB)

Abstract

There are various sources for pollutants in domestic and international waters, among which shipping container accidents releasing microplastics, drilling rigs spilling oil, and accidental release of nucleotide particles are some of the most pressing. To effectively plan disaster responses and assess the long-term effects of such events, particle tracking in natural environments is becoming increasingly relevant. Nu- merous Lagrangian transport simulators exist in the literature; however, these are generally constrained in performance or usability by the use of interpreted or legacy languages, respectively. This project presents a performant Lagrangian simulator developed in C++, designed with an extensible architecture and support for user- accessible Python components. To minimise Python’s runtime overhead, Shedskin, a Python-to-C++ transpiler, is used. Furthermore, the simulator supports a plug- n-play workflow, exchanging select code snippets with user-written implementations at runtime. These plugins are compiled separately from the project, preventing ex- traneous recompilation. In performance evaluation, PanPelagos-Core was 7.1 ± 0.1 times slower than current state-of-the-art simulators, primarily due to its Array- Of-Structure data layout. However, many performance optimisations, including switching to a Structure-Of-Arrays data layout and parallelisation, remain to be implemented.

Item Type: Thesis (Master's Internship Report)
Supervisor name: Kehl, C. and Koellermeier, J.
Degree programme: Computing Science
Thesis type: Master's Internship Report
Language: English
Date Deposited: 07 Jul 2025 09:10
Last Modified: 07 Jul 2025 09:10
URI: https://fse.studenttheses.ub.rug.nl/id/eprint/35888

Actions (login required)

View Item View Item