Open AccessPerformance Optimization of Marine Science and Numerical Modeling on HPC Cluster
Author(s) -
Dongdong Yang,
Hailong Yang,
Luming Wang,
Yucong Zhou,
Zhiyuan Zhang,
Rui Wang,
Yi Liu
Publication year - 2017
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0169130
Subject(s) - speedup , computer science , solver , sensitivity (control systems) , supercomputer , performance prediction , performance improvement , computational science , parameter space , work (physics) , parallel computing , simulation , electronic engineering , mathematics , mechanical engineering , operations management , engineering , programming language , statistics , economics
Marine science and numerical modeling (MASNUM) is widely used in forecasting ocean wave movement, through simulating the variation tendency of the ocean wave. Although efforts have been devoted to improve the performance of MASNUM from various aspects by existing work, there is still large space unexplored for further performance improvement. In this paper, we aim at improving the performance of propagation solver and data access during the simulation, in addition to the efficiency of output I/O and load balance. Our optimizations include several effective techniques such as the algorithm redesign, load distribution optimization, parallel I/O and data access optimization. The experimental results demonstrate that our approach achieves higher performance compared to the state-of-the-art work, about 3.5x speedup without degrading the prediction accuracy. In addition, the parameter sensitivity analysis shows our optimizations are effective under various topography resolutions and output frequencies.