GPU Accelerated Hybrid Particle‐Field Molecular Dynamics: Multi‐Node/Multi‐GPU Implementation and Large‐Scale Benchmarks of the OCCAM Code DOI Creative Commons
Rosario Esposito, Giuseppe Mensitieri, Yan Zhou

et al.

Journal of Computational Chemistry, Journal Year: 2025, Volume and Issue: 46(13)

Published: May 14, 2025

ABSTRACT A parallelization strategy for hybrid particle‐field molecular dynamics (hPF‐MD) simulations on multi‐node multi‐GPU architectures is proposed. Two design principles have been followed to achieve a massively parallel version of the OCCAM code distributed GPU computing: performing all computations only GPUs, minimizing data exchange between CPU and among GPUs. The hPF‐MD scheme particularly suitable develop GPU‐resident low code. Comparison performances obtained using previous multi‐CPU with proposed are reported. Several non‐trivial issues enable applications systems considerable sizes, including large input files handling memory occupation, addressed. Large‐scale benchmarks system sizes up 10 billion particles presented. Performances moderate quantity computational resources highlight feasibility in systematic studies large‐scale multibillion particle systems. This opens possibility perform systematic/routine reveal new insights problems scales previously inaccessible simulations.

Language: Английский

GPU Accelerated Hybrid Particle‐Field Molecular Dynamics: Multi‐Node/Multi‐GPU Implementation and Large‐Scale Benchmarks of the OCCAM Code DOI Creative Commons
Rosario Esposito, Giuseppe Mensitieri, Yan Zhou

et al.

Journal of Computational Chemistry, Journal Year: 2025, Volume and Issue: 46(13)

Published: May 14, 2025

ABSTRACT A parallelization strategy for hybrid particle‐field molecular dynamics (hPF‐MD) simulations on multi‐node multi‐GPU architectures is proposed. Two design principles have been followed to achieve a massively parallel version of the OCCAM code distributed GPU computing: performing all computations only GPUs, minimizing data exchange between CPU and among GPUs. The hPF‐MD scheme particularly suitable develop GPU‐resident low code. Comparison performances obtained using previous multi‐CPU with proposed are reported. Several non‐trivial issues enable applications systems considerable sizes, including large input files handling memory occupation, addressed. Large‐scale benchmarks system sizes up 10 billion particles presented. Performances moderate quantity computational resources highlight feasibility in systematic studies large‐scale multibillion particle systems. This opens possibility perform systematic/routine reveal new insights problems scales previously inaccessible simulations.

Language: Английский

Citations

0