WRF与NPB科学计算程序在云平台上的性能研究  

Performance analysis of WRF and NPB scientific computing programs on scientific cloud platform

作　　者：张心言 任开军[1,2] 赵成武 邓科峰[1,2] ZHANG Xinyan 1, REN Kaijun 1,2 , ZHAO Chengwu 1, DENG Kefeng 1,2(1.College of Computer,National University of Defense Technology, Changsha 410073, China; 2.Academy of Ocean Science and Engineering,National University of Defense Technology, Changsha 410073, Chin)

机构地区：[1]国防科技大学计算机学院,湖南长沙410073 [2]国防科技大学海洋科学与工程研究院,湖南长沙410073

出　　处：《计算机集成制造系统》2018年第7期1816-1821,共6页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金面上资助项目(61572510)~~

摘　　要：为了探究IaaS云平台在运行科学计算程序方面存在的优势与不足,采用虚拟化技术基于高性能计算物理集群创建了NUDT_SCICloud科学云平台,并选择通用NPB标准测试集和包括若干数据与计算流程的大规模气象数值模拟软件WRF模式对该平台进行了全面性能测试,最后将测试结果与物理集群进行了分析对比。实验结果表明,NUDT_SCICloud云平台运行大规模科学计算程序具有可行性,但与物理计算集群相比性能有一定的差距。To explore the advantages and disadvantages of IaaS cloud platform in running scientific computing programs, NUDT_SCICloud scientific cloud platform was constructed based on High Performance Computing（HPC）physical cluster with the help of virtualization technology. A comprehensive performance test of the platform was carried out with both NAS Parallel Benchmark （NPB） standard test set and large-scale meteorological numerical simulation software Weather Research and Forecast Model （WRF） which included a large number of data processing workflow. Furthermore, the test results were compared with those of the physical cluster and were analyzed, which concluded that it was feasible to run large-scale scientific computing programs on NUDT_SCICloud cloud platform, yet a certain performance degradation exits compared with the physical cluster.

关 键 词：云计算 高性能计算 科学云 天气预报模式 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]