中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 计算技术研究所
  3. 中国科学院计算技术研究所
  4. 中国科学院计算技术研究所期刊论文
  5. 英文
An effective communication topology for performance optimization: a case study of the finite-volume wave modeling (FVWAM)

文献类型：期刊论文

作者Pang, Renbo1,2,3; Yu, Fujiang1,2,3; Gao, Yuanyong1; Yuan, Ye1,2,3; Yuan, Liang4; Gao, Zhiyi1
刊名GEOSCIENTIFIC MODEL DEVELOPMENT
出版日期2025-07-08
卷号18期号:13页码:4119-4136
ISSN号1991-959X
DOI10.5194/gmd-18-4119-2025
英文摘要High-resolution models are essential for simulating small-scale processes and topographical features, which play a crucial role in understanding meteorological and oceanic events, as well as climatic patterns. High-resolution modeling requires substantial improvement to the parallel scalability of the model to reduce runtime, while massive parallelism is associated with intensive communications. Point-to-point communication is extensively utilized for neighborhood communication in Earth models due to its flexibility. The distributed graph topology, first introduced in the message-passing interface (MPI) version 2.2, provides a scalable and informative communication method. It has demonstrated significant speedups over the point-to-point communication method based on a variety of synthetic and real-world communication graph datasets. But its application to Earth models for neighborhood communication is rarely studied. In this study, we implemented neighborhood communication using both the traditional point-to-point communication method and the distributed graph communication topology. We then compare their performance in a case study using the finite-volume wave modeling (FVWAM), including both small-scale regional and large-scale global experiments. Based on the Intel MPI library, the distributed graph topology outperforms the point-to-point communication method for inter-node communication with 32 to 32 768 processes. For small-scale regional experiments with 32 to 512 processes, compared to the point-to-point method, the distributed graph topology achieved an average communication time speedup ranging from 4.78 to 11.67. For operational global wave forecasts with 1024 processes, the runtime of the FVWAM was reduced by 40.2 % when the point-to-point communication method was replaced by the distributed graph communication topology.
资助项目National Key Research and Development Program of China[2023YFC3107801]
WOS研究方向Geology
语种英语
WOS记录号WOS:001523714100001
出版者COPERNICUS GESELLSCHAFT MBH
源URL[http://119.78.100.204/handle/2XEOYT63/42014]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Pang, Renbo; Yu, Fujiang; Gao, Yuanyong
作者单位1.Natl Marine Environm Forecasting Ctr, 8 Dahuisi Rd, Beijing 100081, Peoples R China
2.Natl Marine Environm Forecasting Ctr, State Key Lab Satellite Ocean Environm Dynam, 8 Dahuisi Rd, Beijing 100081, Peoples R China
3.Minist Nat Resources China, Key Lab Res Marine Hazard Forecasting, 8 Dahuisi Rd, Beijing 100081, Peoples R China
4.Chinese Acad Sci, Inst Comp Technol, 6 Kexueyuan Nanlu, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Pang, Renbo,Yu, Fujiang,Gao, Yuanyong,et al. An effective communication topology for performance optimization: a case study of the finite-volume wave modeling (FVWAM)[J]. GEOSCIENTIFIC MODEL DEVELOPMENT,2025,18(13):4119-4136.
APA Pang, Renbo,Yu, Fujiang,Gao, Yuanyong,Yuan, Ye,Yuan, Liang,&Gao, Zhiyi.(2025).An effective communication topology for performance optimization: a case study of the finite-volume wave modeling (FVWAM).GEOSCIENTIFIC MODEL DEVELOPMENT,18(13),4119-4136.
MLA Pang, Renbo,et al."An effective communication topology for performance optimization: a case study of the finite-volume wave modeling (FVWAM)".GEOSCIENTIFIC MODEL DEVELOPMENT 18.13(2025):4119-4136.

入库方式： OAI收割

来源：计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。