$π$-PIC: a framework for modular particle-in-cell developments and simulations

Recently proposed modifications of the standard particle-in-cell (PIC) method resolve long-standing limitations such as exact preservation of physically conserved quantities and unbiased ensemble down-sampling. Such advances pave the way for next-generation PIC codes capable of using lower resolution and fewer particles per cell, enabling interactive studies on personal computers and facilitating large-scale parameter scans on supercomputers. Here, we present a Python-controlled framework which promotes the dissemination and adoption of novel PIC developments by providing a unified interface for accommodation, cross-testing, and comparison of PIC solvers and extensions written in Python or low-level languages like C++. To demonstrate the flexibility of the proposed interface, we present and test implementations of several PIC solvers, as well as extensions that are capable of managing absorbing boundaries, moving-window simulations, and tight focusing of laser pulses.

💡 Research Summary

The paper introduces π‑PIC, a modular particle‑in‑cell (PIC) framework designed to streamline the development, testing, and comparison of novel PIC algorithms and physics extensions. Recognizing that existing high‑performance PIC codes such as PIConGPU, Smilei, WarpX, and others excel in specific domains but present steep learning curves for adding new methods, the authors propose a unified interface that couples a Python front‑end with a pre‑compiled, multi‑threaded C++ back‑end via pybind11.

The architecture is organized into three interaction levels. The first level provides a Python API for creating simulation instances, initializing arbitrary particle distributions and electromagnetic fields, advancing the system, and performing runtime read/write of field and particle data. To avoid performance penalties, user‑defined callbacks are decorated with @cfunc, which compiles the Python code into efficient C++ kernels. The second level consists of “extensions,” modular handlers that execute custom operations on fields and particles at every time step. The paper demonstrates three extensions: a minimal‑masking absorbing boundary, a moving‑window that follows high‑speed beams or laser pulses, and a spatial‑to‑focus mapping for tightly focused laser pulses. Extensions can be written in either Python or C++ and are added to the simulation with a single add_handler() call.

The third level encompasses the actual PIC solvers. Each solver is composed of a particle_solver and a field_solver class, allowing independent or concurrent updates of particles and fields. The framework ships with several reference solvers, including a standard FDTD (finite‑difference time‑domain) solver, a spectral solver, and two energy‑conserving solvers (ec and ec2) originally described in Ref.