CRUX：面向GPU集群上基于网格的流体动力学代码的拓扑感知负载均衡器
CRUX: A topology-aware load balancer for mesh-based fluid dynamics codes on GPU clusters
浏览论文内容
中文总结 AI 辅助
针对GPU集群上CFD模拟扩展受网络性能限制的问题，提出拓扑感知负载均衡器CRUX，其性能优于空间填充曲线方法，可支撑下一代天体物理CFD模拟。
中文摘要 AI 辅助
计算能力的快速发展革新了数值模拟，深刻提升了我们对流体和等离子体的理解。计算流体动力学（CFD）模拟在离散网格上求解描述流体或等离子体运动的偏微分方程，是这一进展的核心。近期的进展将传统数值模型的分辨率和运行时间推至前所未有的水平，同时使新代码能够纳入日益复杂的物理过程。然而，这些模拟的进一步扩展已成为重大挑战，很大程度上是因为现代GPU加速集群中，网络能力的提升相较于浮点性能的快速增长相对有限。本文中，我们提出了一种新颖的负载均衡例程CRUX，旨在高效扩展以处理天体物理学中最苛刻的CFD网格。与基于空间填充曲线的传统方法不同，我们的方法动态考虑了采用不同时间步长演化的网格块之间的计算成本差异，同时最小化通信开销，还能考虑异构硬件。通过在OLCF Frontier和ALCF Aurora上使用多达5400个GPU的大量基准测试，我们证明该负载均衡算法在所有关键指标（包括负载均匀性、内存消耗和通信效率）上均优于空间填充曲线方法，是下一代CFD模拟的稳健解决方案。
英文摘要
The rapid growth of computational power has revolutionized numerical simulations, profoundly enhancing our understanding of fluids and plasmas. Computational fluid dynamics (CFD) simulations, which solve partial differential equations governing fluid or plasma motion on discretized grids, have been central to this progress. Recent advances have pushed the resolution and runtime of legacy numerical models to unprecedented levels while enabling newer codes to incorporate increasingly sophisticated physics. However, further scaling of these simulations has become a significant challenge, largely due to the comparatively modest improvements in networking capabilities relative to the rapid growth of floating-point performance in modern GPU-accelerated clusters. In this article, we introduce a novel load-balancing routine CRUX designed to scale efficiently for the most demanding CFD grids in astrophysics. Unlike traditional approaches based on space-filling curves, our method dynamically accounts for computational cost disparities among mesh blocks evolved with different timesteps while minimizing communication overhead. It is also able to take into account heterogeneous hardware. Through an extensive suite of benchmarks featuring up to 5,400 GPUs on OLCF Frontier and ALCF Aurora, we demonstrate that our load-balancing algorithm outperforms space-filling curve methods across all key metrics, including load uniformity, memory consumption, and communication efficiency, making it a robust solution for next-generation CFD simulations.
发表机构
- Center for Relativistic Astrophysics, Georgia Institute of Technology（佐治亚理工学院相对论天体物理中心）
- H. Milton Stewart School of Industrial & Systems Engineering, Georgia Institute of Technology（佐治亚理工学院H.米尔顿·斯图尔特工业与系统工程学校）
机构由 AI 辅助整理，请以论文原文为准。
