A HIP-Compatible Accelerator Backend for Fourier-Bessel Particle-in-Cell Simulations on CPU/DCU Heterogeneous Clusters
Abstract
FBPIC (Fourier-Bessel particle-in-cell) is a high-performance simulation code for relativistic plasma and accelerator physics. Its original accelerator backend relies on Numba CUDA, which limits its direct deployment on accelerators using the HIP (Heterogeneous-Compute Interface for Portability) programming environment, such as DCU (Deep Computing Unit) accelerators. In this work, we develop an accelerator backend compatible with HIP that enables FBPIC to run efficiently on DCU platforms while preserving its Python user interface and high level simulation workflow. For the evaluated LWFA (laser-wakefield acceleration) workloads, the proposed backend achieves 1.32-1.54x speedups over the original FBPIC implementation on an NVIDIA V100 GPU and enables efficient execution on the DCU platform. We also summarize the key lessons learned from porting FBPIC to the DCU platform. Multi-DCU experiments achieve a 1.88x strong-scaling speedup on four accelerators and a 2.72x increase in aggregate throughput at approximately 68\% weak-scaling efficiency, with communication analysis identifying inter-node communication and synchronization as the main scalability limitations. Beyond FBPIC, the proposed approach provides a practical reference for porting and optimizing other scientific computing applications developed with Python on heterogeneous accelerator platforms.
Related
- OPALX: A performance-portable framework for particle accelerator simulations
- Brilliance: toward autonomous computational research in particle accelerator physics
- A particle-in-cell model of beam dynamics in a dielectric wall accelerator
- Isochrone fitting of Galactic globular clusters - IX. Tip of the red-giant branch for 27 clusters and constraints on new particle physics
- Colour-Field Based Particle Categorization for Residual Stress Detection and Reduction in Solid SPH Simulations
BibTeX
@misc{fan2026hipcompatible,
title = {{A HIP-Compatible Accelerator Backend for Fourier-Bessel Particle-in-Cell Simulations on CPU/DCU Heterogeneous Clusters}},
author = {Jingliang Fan and Ruiqing He and Yang Wan and Jiandong Shang and Hengliang Guo and Qiang Chen},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.06680},
eprint = {2609.06680},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.06680},
}
RIS
TY - UNPB
TI - A HIP-Compatible Accelerator Backend for Fourier-Bessel Particle-in-Cell Simulations on CPU/DCU Heterogeneous Clusters
AU - Jingliang Fan
AU - Ruiqing He
AU - Yang Wan
AU - Jiandong Shang
AU - Hengliang Guo
AU - Qiang Chen
PY - 2026
DA - 2026/09/06
PB - arXiv
DO - 10.48550/arxiv.2609.06680
UR - https://arxiv.org/abs/2609.06680
AB - FBPIC (Fourier-Bessel particle-in-cell) is a high-performance simulation code for relativistic plasma and accelerator physics. Its original accelerator backend relies on Numba CUDA, which limits its direct deployment on accelerators using the HIP (Heterogeneous-Compute Interface for Portability) programming environment, such as DCU (Deep Computing Unit) accelerators. In this work, we develop an accelerator backend compatible with HIP that enables FBPIC to run efficiently on DCU platforms while preserving its Python user interface and high level simulation workflow. For the evaluated LWFA (laser-wakefield acceleration) workloads, the proposed backend achieves 1.32-1.54x speedups over the original FBPIC implementation on an NVIDIA V100 GPU and enables efficient execution on the DCU platform. We also summarize the key lessons learned from porting FBPIC to the DCU platform. Multi-DCU experiments achieve a 1.88x strong-scaling speedup on four accelerators and a 2.72x increase in aggregate throughput at approximately 68\% weak-scaling efficiency, with communication analysis identifying inter-node communication and synchronization as the main scalability limitations. Beyond FBPIC, the proposed approach provides a practical reference for porting and optimizing other scientific computing applications developed with Python on heterogeneous accelerator platforms.
ER -