Hybrid EPIC–GOD: An energy–conserving hybrid particle–in–cell code for GPU acceleration using OpenACC

Abstract

Hybrid simulations, which combine ion particles with an inertialess, charge-neutralizing electron fluid, offer a valuable bridge between fully kinetic particle–in–cell (PIC) and magnetohydrodynamic (MHD) approaches. In this paper, we present Hybrid EPIC–GOD, a new hybrid particle–in–cell code designed to ensure strict conservation of both local charge and total energy—two critical properties often neglected in conventional hybrid codes.

Hybrid EPIC–GOD solves the coupled equations governing ion particle dynamics and electromagnetic fields using an iterative scheme. The code exactly satisfies the charge continuity equation, while total energy conservation is guaranteed through the iterative convergence process. We describe the implementation in detail and validate the code's performance across a broad spectrum of plasma processes, including waves, instabilities, collisionless shocks, and magnetic reconnection.

The results show that Hybrid EPIC–GOD accurately reproduces analytical solutions and benchmark results, while maintaining rigorous charge and energy conservation. Moreover, the code is optimized for GPU acceleration using OpenACC, delivering significant performance gains when running on multiple GPUs compared to its CPU–based counterpart. With its combination of accuracy, conservation properties, and computational efficiency, Hybrid EPIC–GOD provides a powerful tool for studying collisionless plasma dynamics in both space and astrophysical environments.