首页> 外文会议>ISCA International Conference on Parallel and Distributed Computing Systems >Bidirectional Division-free Gaussian Elimination Algorithm for Solving Linear Equations
【24h】

Bidirectional Division-free Gaussian Elimination Algorithm for Solving Linear Equations

机译:用于求解线性方程的双向分割高斯消除算法

获取原文

摘要

Multiple-chained divisions and/or square root computations result in an unacceptable cumulative roundoff errors in the solution of linear equations. To circumvent this difficulty, algorithms based on division-free and/or square root-free computations have been proposed. In this paper, a new and faster algorithm based on division-free Gaussian Elimination (GE) for solving a system of linear equations is presented. The new algorithm uses the bidirectional elimination principle proposed in. The mapping of the new algorithm on to systolic arrays is discussed and it is shown that the new algorithm is faster than any of the existing algorithms for solving linear equations. Finally, a problem partitioning technique to solve problems of larger size on an array of smaller size is described.
机译:多链划分和/或平方根计算导致线性方程解中不可接受的累积循环误差。为了避免这种困难,已经提出了基于分割和/或平方无根计算的算法。本文提出了一种基于用于求解线性方程系统的自由高斯消除(GE)的新且更快的算法。新算法使用提出的双向消除原理。讨论了新算法对收缩阵列的映射,并显示新算法比任何用于求解线性方程的现有算法。最后,描述了解决较小尺寸阵列上的较大尺寸问题的问题分配技术。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号