郭庆平
分布式并行处理方向
个性化签名
- 姓名:郭庆平
- 目前身份:
- 担任导师情况:
- 学位:
-
学术头衔:
博士生导师
- 职称:-
-
学科领域:
计算机软件
- 研究兴趣:分布式并行处理方向
郭庆平教授是武汉理工大学的博士生导师。他长期从事分布式并行处理方向的教学与科研工作,为国内和国外在分布式并行处理方向的知名教授。从1986年起,他曾二次经国家教委(国家出国留学基金委)公派,四次由英国皇家学会邀请,到英国伦敦大学玛丽皇后学院、爱丁堡大学并行处理中心、伦敦城市大学、格林威治大学伦敦西敏大学进修学习和合作研究,曾参加英国第五代计算机AVEY 研究计划,其承担的性能评估研究任务曾得到英国同行的好评,并荣获1994年度英国皇家学会的王宽诚奖(K. C. WONG AWARD)。他从1998年1月起到2001年3月止,作为中方负责人主持了一项由英国皇家学会和中国国家自然科学基金委共同资助的中英国际合作项目;从1997年起,连续主持两项国家自然科学基金项目,主持一项省自然科学基金项目。并在2001年应新加坡国立大学计算与数学学院的邀请,去新加坡国立大学计算与数学学院进行并行计算领域的合作研究。郭庆平教授是分布式并行处理国际会议DCABES的发起人、指导委员会成员、DCABES的共同主席之一,曾担任第一届DCABES2001主席、第二届DCABES2002副主席,并正担任第三届DCABES国际会议 (DCABES2004) 的主席; 为DCABES2001国际会议论文集的主编,DCABES2002国际会议论文集的副主编,正担任DCABES2004国际会议论文集的主编。历届DCABES国际会议论文集均被ISTP全部收录。从上述事实可以看出,郭庆平教授有很强的综合素质:他不但业务能力强,而且有很强的组织能力和对外交流能力;他的学术水平高、学风严谨,以身作则,不弄虚作假,在国内外学术界口碑甚佳。他的学术水平高,创新能力强。他在区域分解并行算法设计中首创提出的虚拟边界预测法,大大提高了数值计算中并行算法的加速比,在数值计算区域分解领域开辟了一个新的研究方向,从而能持续得到国家自然科学基金项目的资助郭庆平教授在科研和对外合作交流方面,具有较大的发展潜力;目前他还正在为武汉理工大学计算机科学与技术学院和英国伦敦格林威治大学数学与计算学院洽谈在硕士研究生层次的联合培养、合作办学。郭庆平教授具有优秀的团队精神,领导了一个出色的并行处理研究组。由武汉理工大学校发起、组织和主持的DCABES国际会议,在国内外有较大的影响,为武汉理工大学扩大了国际知名度,加强了我国教育和科技界国内、外的学术合作和交流。
-
主页访问
2662
-
关注数
0
-
成果阅读
350
-
成果数
10
【期刊论文】在多层网络区域分裂并行计算中虚拟边界条件预测算法*1)
郭庆平, 章社生, 卫加宁, Yakup Paker, Dennis Parkinson
数值计算与计算机应用,2000,(4):287~293,-0001,():
-1年11月30日
We propose a virtual boundary condition forecast algorithm for ulti grid paralled computing, and derive a forecast function formula in this paper. Numerical results of one and two-dimension boundary condition problems obtained with the algorithm in a PVM network computing environment show that the algorithm has high parallel effciency.
Multi level,, Multigrid,, Domain decomposition,, Virtual boundary forecast
-
0浏览
-
0点赞
-
0收藏
-
0分享
-
9下载
-
0评论
-
引用
【期刊论文】拓扑重构的交叉开关芯片IMS C004的结构功能及应用*
郭庆平, 郭庆平), 高曙), 李庆华)
武汉交通科技大学学报,1996,20(3):291~294,-0001,():
-1年11月30日
对一种拓扑重构的交叉开关芯片—IMS C004的结构进行了分析,介绍了其使用方法及功能,最后以一个多transputer系统为例,说明了如何利用IMS C004在中等规模的多机系统中构成开关网,从而实现系统的拓扑重构。
拓扑重构, 多处理器系统, 通信延迟
-
0浏览
-
0点赞
-
0收藏
-
0分享
-
5下载
-
0评论
-
引用
郭庆平
华中工学院学报,1987,15(1):59~64,-0001,():
-1年11月30日
本文用频谱图平移迭加的方法导出了模拟信号A/D转拘时周期取样速率应当满足的充分必要条件。文中指出,Nyquist取样定理是此充要条件的一个特例,文献[3]所载的带通信号取样定理也是访充要条件的一个特例。
模拟信号,, 模-数转换器,, 频谱, 带通信号.,
-
0浏览
-
0点赞
-
0收藏
-
0分享
-
6下载
-
0评论
-
引用
郭庆平, Guo Qingping, Wei Jialin, Yakup Paker, Dennis Parkinson, Zhang Sesheng
,-0001,():
-1年11月30日
This paper proposes and theoretically proves a Virtual Boundary Horecast (VBF) multi-grid method for paralled calculation of transient equation. Some forecast theorems. Such as necessary and sufficient comditions of VBF convergence, have been proven with some examples. Using our new method, a quite impressive speed up and higher computer parallel effciency for solving transient equation hanve been obtained.
Domain Decomposition,, Virtual Boundary Forecast,, Multi-grod,, Method,, Network-based Parallel Computing
-
35浏览
-
0点赞
-
0收藏
-
0分享
-
126下载
-
0评论
-
引用
【期刊论文】NETWORK COMPUTING PERDFORMANCE EVALUATION IN MESSAGE PASSING ENVIRONMENT
郭庆平, GUO QINGPING a, *, YAKUP PAKER b, +, DENNIS PARKINSON b, and XIAO JINSHENG a
Paraliel Algoriths and Applioations, 2003 Vol. Not known (not known), pp. 1-8,-0001,():
-1年11月30日
This paper is based on joint rescanch between Queen Mary, University of London, and Wuhan Univesity of Technologys, People's Republic of China, On the use of network conpution for solving non-linera hear flow probles A general formulation for system perfornance evaluation has been derived from measured results, which covers the network womputing as well as multiprocessor computing. The Andahl's law and Gustafon's modification for speedup have been uniformly explained from this formula. Some criterions of speedup, effciency a nd granularity for network computing also have been so ggested.
Network, Message passig environment, PVM, Algorithm
-
53浏览
-
0点赞
-
0收藏
-
0分享
-
112下载
-
0评论
-
引用
【期刊论文】Seismic Processing on TRAM-based Parallel Computers
郭庆平, Guo Qingping, Cai Xiuqing, Qin Youlian, Zhu Tao, Zhu Hongyan Qian Fang, Liu Ping
Transputer Applications and systems 94,-0001,():
-1年11月30日
A parallel system for seismic processing has been recently dveloped in Wuhan Digital Enginessing, and discusses how to build an oill exploration-oriented parallel syste. The architecture, operation environment and its qpplication aspects have been discussed in dectail.
-
44浏览
-
0点赞
-
0收藏
-
0分享
-
57下载
-
0评论
-
引用
郭庆平, Guo Qingping, Dennis Pakfnsor, Xiao forsheng, Yabup Palon
,-0001,():
-1年11月30日
The parallel computing based on the domain deson-position method is used for solving this problem. A mumber of PCs are used to form a network unber the Liux operafing system. A parallel algoxithm of the domain decomposition is fut forvavd by means of the plysical coneption. i.e. the continuous ondition of heat flux in the interace. Another parallel algorthm is Caalled as modiffed tridagrnmal matrix algorithm (mTDMA). In this algraithm, the equation set for a subregion in the domain are elininared to a single equation, and the collection of these equations (ion per slave processor) from a small set of equations. This equation set is solved by using the classical tridiagonal matrix algocithm (TDMA), but by sending and receiving data among the master and slaves. The impliementation of thes domain decomposition paralled algorithm are made in PVM envirment. (to be rewrituen.)
-
54浏览
-
0点赞
-
0收藏
-
0分享
-
64下载
-
0评论
-
引用
郭庆平, GUO Qingping, Yakup Paker, Zhang Shesheng, Dennis Parkinson, and WEI Jianing
J. Comput. Sci. & Technol, 2000, 15 (4): ,-0001,():
-1年11月30日
In this paper, an optimum tacitc of multi-grid parallel algorithm with virtual borundary forecast method is disscussed, and a two-stage implementation is presented. The mumerical pesults of solving a non-optimum one.
algorithm,, parallel multi-grid,, virtual boundary forecast (, VBF), ,, speedup,, network computing,, PVM
-
66浏览
-
0点赞
-
0收藏
-
0分享
-
61下载
-
0评论
-
引用
郭庆平, Guo Qing-ping, and Yakup Paker
Computer Systerm Science & Engineering, 1990, 5 (1): 21~29,-0001,():
-1年11月30日
The parper exposes the concurrency characteristics of the communication between processors on a transputer-based muliprocessor system and proplses on approsimate formula for axxdessing the granlarty of distribution parallel proceses among proessors. The actual measured results for a pipline sorting program running on the Polytechinc of Cenline sortion proessors. The actual measured restdts for a pipline sorting program running on the Polytechin of Central London's transputer machine (the T-rack) confirm these deduced results.
transputer,, parallel processing,, pipeline sorting,, concurrency,, granularity
-
60浏览
-
0点赞
-
0收藏
-
0分享
-
68下载
-
0评论
-
引用
郭庆平, Guo Qing-ping, Zhang Shesheng, Yakup Paker, Denmis Parkinson, Wei Jialin
,-0001,():
-1年11月30日
We present convergence moldel for parallel multi-grod sclution of non-linear transien tequations using the virtual boundary forecast (VBF) method, and derive a relation. For function errors in whole domain and sub-domains, Numerical convergence theorems are deseribed and proved, and iterative loop thimes estimated. The results of a non-linear transient heat transfer problem solution with this method show that It has a very fast-convergence rate.
Multi-Gird,, Domain Decomposition,, Math-ematics Model Parallel Processing,, Virtual Boundary Gondition Forecast
-
38浏览
-
0点赞
-
0收藏
-
0分享
-
76下载
-
0评论
-
引用