• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2012, Vol. 34 ›› Issue (2): 87-92.

• 论文 • 上一篇    下一篇

面向集合预报的高性能计算环境

刘灿灿,张卫民,骆志刚,任开军   

  1. (国防科学技术大学计算机学院,湖南 长沙 410073)
  • 收稿日期:2010-01-28 修回日期:2010-05-18 出版日期:2012-02-25 发布日期:2012-02-25

HighPerformance Computing Environments for Ensemble Prediction

LIU Cancan,ZHANG Weimin,LUO Zhigang,REN Kaijun   

  1. (School of Computer Science,National University of Defense Technology,Changsha 410073,China)
  • Received:2010-01-28 Revised:2010-05-18 Online:2012-02-25 Published:2012-02-25

摘要:

集合预报中需要大量的高性能计算资源对海量数据进行实时分析和处理,高效的资源管理和数据共享将有效提高预报的效率和时效性。本文在分析集合预报的特点和需求的基础上,设计了基于元数据提取的海量数据管理方案和基于虚拟组织的高性能计算资源管理方案,并采用网格技术对这些资源进行有效管理和共享,为分布在各个地域、不同组织的气象科学家提供一个高效共享的协同开发平台,达到有效提高预报结果的时效性并推动中尺度天气预报事业发展的目标。

关键词: 中尺度天气事件, 集合预报, 高性能计算, 网格计算, 元数据

Abstract:

There is an urgent demand for High Performance Computing (HPC) resources to analyze and process huge data in ensemble prediction. It will improve the effect of the prediction result to provide effective resources and data management. With the huge data and high demand for HPCs, we provide a data management strategy based on meta data distillation and a resource management method based on virtual organization. Besides, the grid technology is used for data and resource sharing and management. As a result, a coordinated development environment is built for the meteorologists distributed all over the world and different organizations, which accelerates the development of meteorology as a result.

Key words: mesoscale weather events;ensemble prediction;high performance computing(HPC);grid computing;meta data