The International Conference for High Performance Computing, Networking, Storage, and Analysis
The International Conference for High Performance Computing, Networking, Storage, and Analysis
Authors: Yumeng Shi (Computer Network Information Center, Chinese Academy of Sciences); Ningming Nie (Computer Network Information Center, Chinese Academy of Sciences; University of Chinese Academy of Sciences); Shunde Li and Jue Wang (Computer Network Information Center, Chinese Academy of Sciences); Kehao Lin (Hangzhou Dianzi University; Computer Network Information Center, Chinese Academy of Sciences); Chunbao Zhou (Computer Network Information Center, Chinese Academy of Sciences); Shigang Li (School of Computer Science, Beijing University of Posts and Telecommunications); Kehan Yao (Hangzhou Dianzi University); Yangde Feng (Computer Network Information Center, Chinese Academy of Sciences); Yan Zeng (Hangzhou Dianzi University); Fang Liu and Yangang Wang (Computer Network Information Center, Chinese Academy of Sciences); and Yue Gao (China Institute of Atomic Energy)
Abstract: Structural dynamics simulation plays an important role in research on reactor design and complex engineering. The Hybrid Total Finite Element Tearing and Interconnecting (HTFETI) method combined with Newmark method is an efficient way to solve large-scale structural dynamics problems. However, the sparse direct solver and the load imbalance caused by inconsistent density models are two critical issues limiting the performance and the scalability of structural dynamics computing. For the former, we propose an efficient variable-size batched method to accelerate SpMV on GPUs. For the latter, we establish an online performance prediction model, based on which we then design a novel inter-cluster subdomain fine-tuning algorithm to balance the workload of HTFETI parallel computing. We are the first to achieve the high-fidelity structural dynamics simulation of China Experimental Fast Reactor core assembly with up to 53.4 billion grids. The weak and strong scalability efficiencies reach 91.77% and 86.13% on 12,800 GPUs, respectively.
Back to Technical Papers Archive Listing