Accelerating the Particle-In-Cell code ECsim with OpenACC

2026-03-17

AI Summary

This work presents the first integration of OpenACC into the semi-implicit, energy-conserving particle-in-cell code ECsim, targeting exascale supercomputing with minimal code modifications to achieve lightweight GPU acceleration. By leveraging the unified memory architecture of GH200 and applying optimizations for both strong and weak scalability, the approach substantially enhances performance and energy efficiency. Evaluated on the Leonardo system, the accelerated implementation achieves a 5× speedup and reduces energy consumption by a factor of three compared to the CPU-only version. Strong scaling efficiency reaches 70% on 64 GPUs, while weak scaling efficiency attains 78% on 1,024 GPUs, demonstrating the solution’s effectiveness at scale.

Technology Category

• Machine Learning: Hardware-aware ML
• Search and Optimization: Sampling/Simulation-based Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The Particle-In-Cell (PIC) method is a computational technique widely used in plasma physics to model plasmas at the kinetic level. In this work, we present our effort to prepare the semi-implicit energy-conserving PIC code ECsim for exascale architectures. To achieve this, we adopted a pragma-based acceleration strategy using OpenACC, which enables high performance while requiring minimal code restructuring. On the pre-exascale Leonardo system, the accelerated code achieves a $5 \times$ speedup and a $3 \times$ reduction in energy consumption compared to the CPU reference code. Performance comparisons across multiple NVIDIA GPU generations show substantial benefits from the GH200 unified memory architecture. Finally, strong and weak scaling tests on Leonardo demonstrate efficiency of $70 \%$ and $78 \%$ up to 64 and 1024 GPUs, respectively.

Problem

• Particle-In-Cell
• ECsim
• exascale
• code acceleration
• high-performance computing

Innovation

• OpenACC
• Particle-In-Cell
• GPU acceleration
• exascale computing
• energy efficiency