Skip to main navigation Skip to search Skip to main content

Improving the Parallelism of CESM on GPU

  • Beihang University
  • Fourth Research Institute of Telecommunications Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Community Earth System Model (CESM) is one of the most popular climatology research models. However, the computation of CESM is quite expensive and usually lasts for weeks even on high-performance clusters. In this paper, we propose several optimization strategies to improve the parallelism of three hotspots in CESM on GPU. Specifically, we analyze the performance bottleneck of CESM and propose corresponding GPU accelerations. The experiment results show that after applying our GPU optimizations, the kernels of the physical model achieve significant performance speedup respectively.

Original languageEnglish
Title of host publicationAlgorithms and Architectures for Parallel Processing - 19th International Conference, ICA3PP 2019, Proceedings
EditorsSheng Wen, Albert Zomaya, Laurence T. Yang
PublisherSpringer
Pages11-18
Number of pages8
ISBN (Print)9783030389604
DOIs
StatePublished - 2020
Event19th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2019 - Melbourne, Australia
Duration: 9 Dec 201911 Dec 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11945 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2019
Country/TerritoryAustralia
CityMelbourne
Period9/12/1911/12/19

Keywords

  • CESM
  • GPU
  • Performance optimization

Fingerprint

Dive into the research topics of 'Improving the Parallelism of CESM on GPU'. Together they form a unique fingerprint.

Cite this