swNEMO_v4.0: an ocean model based on NEMO4 for the new-generation Sunway supercomputer

Y. Ye; Y. Ye; Y. Ye; Y. Ye; Z. Song; Z. Song; Z. Song; S. Zhou; S. Zhou; S. Zhou; Y. Liu; Q. Shu; Q. Shu; Q. Shu; B. Wang; W. Liu; W. Liu; F. Qiao; F. Qiao; F. Qiao; L. Wang; L. Wang

doi:10.5194/gmd-15-5739-2022

Geoscientific Model Development (Jul 2022)

swNEMO_v4.0: an ocean model based on NEMO4 for the new-generation Sunway supercomputer

Y. Ye,
Y. Ye,
Y. Ye,
Y. Ye,
Z. Song,
Z. Song,
Z. Song,
S. Zhou,
S. Zhou,
S. Zhou,
Y. Liu,
Q. Shu,
Q. Shu,
Q. Shu,
B. Wang,
W. Liu,
W. Liu,
F. Qiao,
F. Qiao,
F. Qiao,
L. Wang,
L. Wang

Affiliations

Y. Ye: National Supercomputing Center, Wuxi 214000, China
Y. Ye: First Institute of Oceanography, and Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China
Y. Ye: Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China
Y. Ye: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
Z. Song: First Institute of Oceanography, and Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China
Z. Song: Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China
Z. Song: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
S. Zhou: First Institute of Oceanography, and Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China
S. Zhou: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
S. Zhou: School of Software, Shandong University, Jinan 250101, China
Y. Liu: School of Data Science and Engineering, East China Normal University, Shanghai 200062, China
Q. Shu: First Institute of Oceanography, and Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China
Q. Shu: Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China
Q. Shu: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
B. Wang: National Supercomputing Center, Wuxi 214000, China
W. Liu: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
W. Liu: School of Software, Shandong University, Jinan 250101, China
F. Qiao: First Institute of Oceanography, and Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China
F. Qiao: Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China
F. Qiao: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
L. Wang: Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology, Qingdao 266237, China
L. Wang: College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China

DOI: https://doi.org/10.5194/gmd-15-5739-2022
Journal volume & issue: Vol. 15
pp. 5739 – 5756

Abstract

Read online

The current large-scale parallel barrier of ocean general circulation models (OGCMs) makes it difficult to meet the computing demand of high resolution. Fully considering both the computational characteristics of OGCMs and the heterogeneous many-core architecture of the new Sunway supercomputer, swNEMO_v4.0, based on NEMO4 (Nucleus for European Modelling of the Ocean version 4), is developed with ultrahigh scalability. Three innovations and breakthroughs are shown in our work: (1) a highly adaptive, efficient four-level parallelization framework for OGCMs is proposed to release a new level of parallelism along the compute-dependency column dimension. (2) A many-core optimization method using blocking by remote memory access (RMA) and a dynamic cache scheduling strategy is applied, effectively utilizing the temporal and spatial locality of data. The test shows that the actual direct memory access (DMA) bandwidth is greater than 90 % of the ideal band-width after optimization, and the maximum is up to 95 %. (3) A mixed-precision optimization method with half, single and double precision is explored, which can effectively improve the computation performance while maintaining the simulated accuracy of OGCMs. The results demonstrate that swNEMO_v4.0 has ultrahigh scalability, achieving up to 99.29 % parallel efficiency with a resolution of 500 m using 27 988 480 cores, reaching the peak performance with 1.97 PFLOPS.

Published in Geoscientific Model Development

ISSN: 1991-959X (Print); 1991-9603 (Online)
Publisher: Copernicus Publications
Country of publisher: Germany
LCC subjects: Science: Geology
Website: https://www.geoscientific-model-development.net/

About the journal