Single Domain Multiple Decompositions for Particle-in-Cell simulations

TL;DR

This paper introduces the SDMD method for Particle-in-Cell simulations, enabling different domain decompositions for particles and fields to improve load balancing, parallelism, and compatibility with spectral solvers.

Contribution

The paper presents the SDMD approach, allowing simultaneous use of multiple domain decompositions in PIC codes to address compatibility and performance challenges.

Findings

Improved load balancing through small particle domains.

Reduced synchronization costs in PIC simulations.

Enhanced parallelism and spectral solver compatibility.

Abstract

As a multi-purpose Particle-In-Cell (PIC) code, Smilei gathers many different features in a single software. Combining some of them is challenging. In particular, spectral solvers and patch based load balancing have a priori non compatible requirements. This paper introduces the Single Domain Multiple Decompositions (SDMD) method in order to address this issue. To do so, different domain decompositions are used for fields and particles operations. This approach allows to keep small domains for particles, necessary for a good load balancing, while having large domains for the fields. It proves beneficial in mitigating synchronization costs and gives the opportunity to introduce more paralellism in the PIC algorithm on top of providing structures compatible with spectral solvers.