首页 | 本学科首页   官方微博 | 高级检索  
     

数据相关性分析中替代约束的有效性
引用本文:乔林,黄维通,孟威,汤志忠. 数据相关性分析中替代约束的有效性[J]. 计算机工程与设计, 2006, 27(5): 724-727
作者姓名:乔林  黄维通  孟威  汤志忠
作者单位:清华大学,计算机与信息管理中心,北京,100084;清华大学,计算机系高性能计算所,北京,100084
摘    要:数据相关性分析经常使用诸如约束有效性的几个测度,详细分析了什等式约束要比方向向量约束更有效,体性测试算法。方向向量或体差不等式等作为替代约束以加强求解丢番图方程的能力。给出了替代么样的替代约束是可行的、有效的、良性的或恰当的,并证明在大多数情况下体差不差不等式测试方法的收敛性要强于其它使用方向向量作为替代约束的传统数据相关性测试算法。

关 键 词:软件流水  数据相关性  替代约束  迭代向量  体差不等式测试  有效约束
文章编号:1000-7024(2006)05-0724-04
收稿时间:2005-03-12
修稿时间:2005-03-12

Effectiveness of surrogating constraints in data dependence analysis
QIAO Lin,HUANG Wei-tong,MENG Wei,TANG Zhi-zhong. Effectiveness of surrogating constraints in data dependence analysis[J]. Computer Engineering and Design, 2006, 27(5): 724-727
Authors:QIAO Lin  HUANG Wei-tong  MENG Wei  TANG Zhi-zhong
Affiliation:1. Computer and Information Management Center, Tsinghua University, Beijing 100084, China; 2. Institute of High Performance Computing, Department of Computer Science, Tsinghua University, Beijing 100084, China
Abstract:Most data dependence analysis methods often use surrogating constraints, e.g., direction vectors and dependence difference inequalities, to enforce their capabilities for solving the Diophantine equation. Several measures of effectiveness of surrogating constraints were presented, what surrogating constraints are feasible, effective, well and exact was discussed in detail. In most cases, as the paper proved, dependence difference inequalities arc more effective than direction vectors, and this leads to that the convergence of dependence difference inequality test is better than other traditional data dependence analysis methods by using direction vectors as surrogating con- straints.
Keywords:software pipelining   data dependence   surrogating constraint   iteration vector   dependence difference inequality test   effective constraint
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