一种加速大规模线性方程组求解的并行方法

作者：吴丹红 日期：2008-04-23/span> 浏览：3923 查看PDF文档

一种加速大规模线性方程组求解的并行方法
吴丹红
（杭州电子科技大学 计算机智能与软件技术研究所，浙江 杭州 310018）

摘 要：在计算机辅助几何造型、计算机辅助决策、计算机数值计算等领域，需要求解大规模的线性方程组，对此类线性方程组的求解是一个耗时极其巨大的过程。根据计算机多核CPU的最新发展技术，从求解线性方程组的LU分解算法入手，提出了一种加速求解大规模线性方程组的并行计算方法。实验证明：该并行优化算法能较好地提高线性方程组的计算速度。
关键词：LU分解;并行处理;双核

Parallel accelerating method of solving largescale linear equations
WU Danhong
(Institute of Intelligent and Software Technology, Hangzhou Dianzi University, Hangzhou 310018, China)

Abstract:Solving large scale linear equations is necessary in the areas of computer aided geometric modeling, computer aided decision, computer numerical computing and so on. Such linear equation solving is a hugely timeconsuming process. According to the computer, the latest developments in multicore CPU technology, a parallel accelerating method of solving largescale linear equations based on the traditional LU decompression algorithm of solving linear equations was presented. The experiments indicate that the method is able to improve the speed of solving large scale linear equations excellently.
Key words: LU decompression; parallel computing; dual core CPU