Neutron particle transport 3D method of characteristic Multi GPU platform Parallel Computing

TL;DR

3D neutron transport with the Method of Characteristics (MOC) is accurate but computationally demanding; this work implements a GPU-based parallel MOC using On-The-Fly ray tracing to generate and process characteristic lines on-the-fly, paired with line preloading and load balancing for scalability. Key contributions include a portable CUDA/HIP implementation, a Z-STACK memory layout for efficient data access, and a preloading plus load-balancing strategy that yields large-scale speedups while preserving accuracy. Validation against the OECD/NEA C5G7 3D benchmark confirms identical $k_{eff}$ and iteration counts, while parallel tests show GPU speedups from $30$–$100$× as problem size increases. The approach enables high-precision, large-scale deterministic neutron transport on multi-GPU platforms and lays groundwork for extensions to additional benchmarks and matrix-based representations to further boost parallel efficiency.

Abstract

Three-dimensional neutron transport calculations using the Method of Characteristics (MOC) are highly regarded for their exceptional computational efficiency, precision, and stability. Nevertheless, when dealing with extensive-scale computations, the computational demands are substantial, leading to prolonged computation times. To address this challenge while considering GPU memory limitations, this study transplants the real-time generation and characteristic line computation techniques onto the GPU platform. Empirical evidence emphasizes that the GPU-optimized approach maintains a heightened level of precision in computation results and produces a significant acceleration effect. Furthermore, to fully harness the computational capabilities of GPUs, a dual approach involving characteristic line preloading and load balancing mechanisms is adopted, further enhancing computational efficiency. The resulting increase in computational efficiency, compared to traditional methods, reaches an impressive 300 to 400-fold improvement.

Figures (8)

• Figure 1: Ray Tracing Track Z-Stacks
• Figure 2: Overview of Functional Modules
• Figure 3: Flowchart of Kernel Function
• Figure 4: struct of Z-STACK
• Figure 5: Overview of Track Manangment