Towards a complete task-based implementation of a 3D particle-in-cell code: Performance studies and benchmarks

• Published in: Computer physics communications Vol. 313; p. 109647
• Main Authors: Silva-Cuevas, J.J., Zych, M., Peyen, K., Kabadshow, I., Lobet, M.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2025
• Subjects: Kinetic plasma simulations; Parallel computation; Particle-in-cell algorithms; Plasma modeling; Task-based asynchronous methods; Parallel computation; Kinetic plasma simulations; Task-based asynchronous methods; Plasma modeling; Particle-in-cell algorithms
• ISSN: 0010-4655
• DOI: 10.1016/j.cpc.2025.109647

This article investigates different programming models for miniPIC, a Particle-In-Cell mini-app, to improve overall scalability. An innovative implementation of a fully asynchronous task-based implementation of a 3D Particle-In-Cell code miniPIC is presented for the first time. The task-based model has been specially implemented in the particle-in-cell code via the backend OpenMP and the library Eventify. Four physical studies were selected as benchmarks: thermal plasma, plasma beam diffusion, a laser colliding with a plasma beam sphere, and thermal plasma with an imbalanced artificial operator. Besides, different parametric studies were designed to measure the scalability of the implementation to varying numbers of cores and various physical conditions. The current parametric studies were performed in an Intel cascade lake-based machine and an AMD EPYC-based machine to measure scalability performance on different architectures.