首页 | 本学科首页   官方微博 | 高级检索  
     

软件流水领域二维数组的体差不等式测试算法
引用本文:乔林,黄维通,孟威,汤志忠.软件流水领域二维数组的体差不等式测试算法[J].计算机工程,2005,31(21):1-3,24.
作者姓名:乔林  黄维通  孟威  汤志忠
作者单位:1. 清华大学计算中心,北京,100084
2. 清华大学计算机科学与技术系高性能计算所,北京,100084
基金项目:国家自然科学基金资助项目(69773028,60173010)
摘    要:给出了二维数组的体差不等式测试算法,并证明二维数组的体差不等式测试算法具有多项式时间复杂度。相对于直接求解,体差不等式测试算法的时间复杂度要小得多。实验表明,对于科学计算循环中出现的大多数数据相关性判定问题,使用体差不等式测试算法可以获得很好的结果。

关 键 词:数据相关性  二维数组  迭代向量  体差不等式测试  可行解域
文章编号:1000-3428(2005)21-0001-03
收稿时间:2004-12-01
修稿时间:2004-12-01

Dependence Difference Inequality Test Algorithm for Two-dimensional Arrays in Software Pipelining Field
QIAO Lin,HUANG Weitong,MENG Wei,TANG Zhizhong.Dependence Difference Inequality Test Algorithm for Two-dimensional Arrays in Software Pipelining Field[J].Computer Engineering,2005,31(21):1-3,24.
Authors:QIAO Lin  HUANG Weitong  MENG Wei  TANG Zhizhong
Affiliation:1. Computer and Information Managemenl Center, Tsinghua University. Beijing 100084; 2. Institute of High Performance Computing, Department of Computer Science and Technology, Tsinghua University, Beijing 100084
Abstract:This paper presents a new dependence difference inequality test algorithm for two-dimensional arrays, and proves that the time complexity of the algorithm is polynomial. Compared with directly solving algorithms, the time complexity of the dependence difference inequality test algorithm is much better than those of other algorithms. Experimental results also show that the dependence difference inequality test algorithm works well for most of data dependencies appeared in scientific computation programs.
Keywords:Data dependence  Two-dimensional array  lteration vector  Dependence difference inequality test  Feasible solution region
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号