计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (6): 46-50.DOI: 10.3778/j.issn.1002-8331.1508-0214

• 理论与研发 • 上一篇    下一篇

气候系统模式不确定分析工作流执行平台

高  健1,2,3,薛  巍1,2,3,杨广文1,2,3   

  1. 1.清华大学 计算机科学与技术系,北京 100084
    2.清华大学 信息科学与技术国家实验室(筹),北京 100084
    3.清华大学 地球系统数值模拟教育部重点实验室,地球系统科学研究中心,北京 100084
  • 出版日期:2017-03-15 发布日期:2017-05-11

Workflow execution platform for uncertainty analysis of climate system model

GAO Jian1,2,3, XUE Wei1,2,3, YANG Guangwen1,2,3   

  1. 1.Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China
    2.National Laboratory for Information Science and Technology, Tsinghua University, Beijing 100084, China
    3.Ministry of Education Key Laboratory for Earth System Modeling, Center for Earth System Science, Tsinghua University, Beijing 100084, China
  • Online:2017-03-15 Published:2017-05-11

摘要: 气候系统模式的不确定性量化分析包含复杂的工作流执行流程。对其中量化分析过程和模式后处理任务两个阶段的工作流进行了分析,设计和实现了针对该分析工作的工作流执行平台。该平台能够自动化执行典型的气候系统仿真实验与分析的工作流程,支持通过专家知识灵活配置工作流。平台中包含工作流层面的容错设计,支持通过分布式并发执行来加速工作流中的大计算量任务。同时,平台支持用户自定义插件的接入,增强了工作流系统的模块化和规范化。使用设计的平台对GAMIL2气候系统模式进行了物理过程参数不确定性的分析实验,验证了本平台的可行性和有效性。

关键词: 气候系统模式, 不确定性量化分析, 工作流执行, 模块化

Abstract: Quantitative analysis of the uncertainty in the climate system model includes a complex workflow execution process. In this paper, the workflow of the two stages of the quantitative analysis process and the model post processing task is analyzed, and a targeted workflow execution platform is designed and implemented. The platform is capable of automatically performing the workflow of typical climate system simulation experiment and analysis, and supports the flexible configuration of workflow through expert knowledge. Fault tolerant design of the workflow layer is included in the platform, which supports acceleration of the highly computational tasks in the workflow via distributed concurrent execution. At the same time, the platform supports user defined plug-ins, and enhances the workflow system’s modularization and standardization. This paper performs an analysis experiment on physical process parameter uncertainty of the GAMIL2 climate system model using the designed platform. The feasibility and effectiveness of the platform are verified.

Key words: climate system model, uncertainty quantification, workflow execution, modularity