Kilometer-Level Coupled Modeling Using 40 Million Cores: An Eight-Year Journey of Model Development
Abstract
With current and future leading systems adopting heterogeneous architectures, adapting existing models for heterogeneous supercomputers is of urgent need for improving model resolution and reducing modeling uncertainty. This paper presents our three-week effort on porting a complex earth system model, CESM 2.2, to a 40-million-core Sunway supercomputer. Taking a non-intrusive approach that tries to minimizes manual code modifications, our project tries to achieve both improvement of performance and consistency of the model code. By using a hierarchical grid system and an OpenMP-based offloading toolkit, our porting and parallelization effort covers over 80% of the code, and achieves a simulation speed of 340 SDPD (simulated days per day) for 5-km atmosphere, 265 SDPD for 3-km ocean, and 222 SDPD for a coupled model, thus making multi-year or even multi-decadal experiments at such high resolution possible.
Keywords
Cite
@article{arxiv.2404.10253,
title = {Kilometer-Level Coupled Modeling Using 40 Million Cores: An Eight-Year Journey of Model Development},
author = {Xiaohui Duan and Yuxuan Li and Zhao Liu and Bin Yang and Juepeng Zheng and Haohuan Fu and Shaoqing Zhang and Shiming Xu and Yang Gao and Wei Xue and Di Wei and Xiaojing Lv and Lifeng Yan and Haopeng Huang and Haitian Lu and Lingfeng Wan and Haoran Lin and Qixin Chang and Chenlin Li and Quanjie He and Zeyu Song and Xuantong Wang and Yangyang Yu and Xilong Fan and Zhaopeng Qu and Yankun Xu and Xiuwen Guo and Yunlong Fei and Zhaoying Wang and Mingkui Li and Yingjing Jiang and Lv Lu and Liang Su and Jiayu Fu and Peinan Yu and Weiguo Liu and Lixin Wu and Lanning Wang and Xin Liu and Dexun Chen and Guangwen Yang},
journal= {arXiv preprint arXiv:2404.10253},
year = {2024}
}
Comments
18 pages, 13 figures