并行集群系统的Linpack性能测试分析  

ANALYSIS OF LINPACK PERFORMANCE TEST ON PARALLEL CLUSTER SYSTEM

作　　者：罗水华 杨广文[1] 张林波[2] 石威[1] 郑纬民[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084 [2]中国科学院计算数学与科学工程计算研究所,北京100080

出　　处：《数值计算与计算机应用》2003年第4期285-292,共8页Journal on Numerical Methods and Computer Applications

基　　金：国家自然科学基金(60173007);国家863高科技项目基金(2001AA111080;2002AA104580)

摘　　要：§1.引言 近些年随着计算机软硬件技术的提高,尤其是网络部件性能的提高,集群技术得到不断的发展.传统的PVP(Parallel Vector Processor)超级计算机以及MPP(Massively Parallel Processing)的成本很容易达到几千万美元,与此相比,具有相同峰值性能的机群价格则要低1到2个数量级.A fast and accurate method for testing the float-point performance on parallel systems has been proposed by using the HPL benchmark on a cluster system connected by Myrinet network. It is found that HPL shows very good scalability for different BLAS implementations on this system. The factors which mostly affect the result of performance test are: BLAS, array of processors, block size of LU factorization and size of linear system etc. We also found that higher performance can be achieved by using shared memory entirely for communication on each node.

关 键 词：并行集群系统 Linpack性能 Moore定律 计算机系统 性能测试 

分 类 号：TP338[自动化与计算机技术—计算机系统结构]