气候模式模拟中海量数据的并行抽取与可视化分析诊断

地震 ›› 2013, Vol. 33 ›› Issue (4): 214-224.

气候模式模拟中海量数据的并行抽取与可视化分析诊断

邓春林, 张怀, 谭清海, 王蕊, 郭晓虎

中国科学院计算地球动力学重点实验室, 北京 100049

收稿日期:2013-01-11 修回日期:2013-03-19 发布日期:2020-09-27
作者简介:邓春林(1988-), 男, 四川蓬安人, 研究生, 主要从事地学数据可视化、地球动力学等研究。
基金资助:
国家高技术研究发展计划重点项目(2010AA012402); SinoProbe-07-01(0841042G02)

Parallel Extraction and Visualization Analysis of Massive Data Produced by Climate Model Numerical Simulation

DENG Chun-lin, ZHANG Huai, TAN Qing-hai, WANG Rui, GUO Xiao-hu

Laboratory of Computational Geodynamics, CAS, Beijing 100049, China

Received:2013-01-11 Revised:2013-03-19 Published:2020-09-27

摘要/Abstract

摘要： 在地球动力学和气候模拟等领域, 数值模拟产生的数据规模达到Tb至Pb量级。实现这些海量数据的实时可视化和实时诊断分析面临很大的困难, 时空多尺度数据抽取可以解决这一瓶颈。高精度数据的可视化结果展示需要高分辨率的显示设备, 并行的大屏幕显示技术是解决这一问题的有效手段。地球科学数值模拟、并行数据抽取和高分辨率显示都需要搭建高性能计算机集群。本文在基于Lagrange插值的多维度、多尺度、多分辨率并行数据抽取算法的同时, 利用并行计算节点及LCD显示器, 基于Rocks cluster系统搭建起一个176核, 4×10×1024×1280分辨率的高性能计算模拟、数据抽取和并行显示输出的集成平台, 并将该平台成功应用到气候模式模拟产生的海量数据的并行抽取和并行显示。

关键词: 海量数据, Lagrange插值, 数据抽取, 并行可视化, 显示墙

Abstract: The scale of data produced by numerical simulation have reached Tb to Pb magnitude in the fields of geodynamics and climate simulation. It is very difficulty to realize real-time visualization and analysis for this massive data. Data extraction of multi-scale for space and time dimensions is able to solve it. The visualization products of high-precision need high-resolution display device, and parallel display array with a high- resolution can meet this need. Numerical simulation of geosciences, parallel abstract for massive data and high-resolution display all need to construct high-performance computer cluster. We achieved parallel abstract of multi-dimension, multi-scale and multi-resolution program for massive data based on Lagrange interpolation. At the same time, on the base of the Rockscluster system, we used the parallel compute node and the LCD display hardware to build a 4×10 parallel display array with a 52 millions resolution and 176 processors integration platform, which can be used to compute and parallel display. We successfully realized parallel abstract and displayed for massive data produced by climate model numerical simulation using this platform.

Key words: Massive data, Lagrange interpolation, Data extract, Parallel visualization

中图分类号:

P315

邓春林, 张怀, 谭清海, 王蕊, 郭晓虎. 气候模式模拟中海量数据的并行抽取与可视化分析诊断[J]. 地震, 2013, 33(4): 214-224.

DENG Chun-lin, ZHANG Huai, TAN Qing-hai, WANG Rui, GUO Xiao-hu. Parallel Extraction and Visualization Analysis of Massive Data Produced by Climate Model Numerical Simulation[J]. EARTHQUAKE, 2013, 33(4): 214-224.

参考文献

[1] Richard L Burden. 数值分析(第七版翻译版)[M]. 北京: 高等教育出版社, 2005.
[2] 陈国良. 并行算法实践[M]. 北京: 高等教育出版社, 2004.
[3] USGS[OL]: http://earthquake.usgs.gov/earthquakes/eqinthenews/2011.
[4] VTK[OL]: http://public.kitware.com/VTK.
[5] 陈绍林, 张怀, 陈石, 等. 可扩展大屏幕高分辨率并行显示系统的构建及其在地学中的应用[J]. 中国科学院研究生院学报, 2009, 26(2): 243-250.
[6] 陈国良. 并行计算机体系结构[M]. 北京: 高等教育出版社, 2002.
[7] Martin Kevin E. DMX.[2006-12-07][OL]. http://dmx.sourceforge.net
[8] Renambot L, Rao A, Singh R, et al. Sage: the scalable adaptive graphics environment[J]//Proceedings of WACE. 2004, 9(23): 2004-09.
[9] Doerr K, Kuester F. CGLX: a scalable, high-performance visualization framework for networked display environments[J]. Visualization and Computer Graphics, IEEE Transactions on, 2011, 17(3): 320-332.
[10] Humphreys G, Houston M, Ng R, et al. Chromium: a stream-processing framework for interactive rendering on clusters[J]//ACM Transactions on Graphics (TOG). ACM, 2002, 21(3): 693-702.
[11] Xinerama[OL]http://linuxdocs.org/HOWTOs/Xinerama-HOWTO.html
[12] Hamano T, Fujihara R, Yamashita A. DEVELOPMENT OF LARGE, HIGH-RESOLUTION DISPLAY FOR SPRING-8 CENTRAL CONTROL ROOM[J].
[13] Humphreys G, Eldridge M, Buck I, et al. WireGL: a scalable graphics system for clusters[A]//International Conference on Computer Graphics and Interactive Techniques: Proceedings of the 28 th annual conference on Computer graphics and interactive techniques[C]. 2001. 129-140.
[14] Rocks cluster[OL]: http://www.rocksclusters.org/
[15] HPC[OL]: http://central6.rocksclusters.org/roll-documentation/hpc
[16] VizRoll[OL]: http://www.rocksclusters.org/roll-documentation/viz/5.4/
[17] PVFS[OL]: http://www.pvfs.org/
[18] Ganglia[OL]: http://central6.rocksclusters.org/roll-documentation/ ganglia
[19] SGE[OL]: http://central6.rocksclusters.org/roll-documentation/sge
[20] Web-server[OL]: http://en.wikipedia.org/wiki/Web_server
[21] Base[OL]: http://central6.rocksclusters.org/roll-documentation/base
[22] Kernel[OL]: http://www.kernel.org/
[23] OS[OL]: http://www.centos.org/
[24] Google Earth[OL]: http://earth.google.com