期刊文献+

稀疏有限元线性系统的并行算法实现

Parallel algorithm for sparse finite element linear systems
下载PDF
导出
摘要 在对称多处理机系统上,提出了一种求解稀疏对称有限元线性系统的正规化精确并行逆算法。该算法以一种避免数据依赖的反对角运动方法为基础,使用OpenMP编译指导来实现。诸如加速比和效率等数值实验结果的推出,说明在一个对称多处理机系统上,所提出的算法求解方法能更好地提高性能,获得更大的加速。 A new parallel normalized exact inverse algorithm is presented for solving sparse symmetric finite element linear systems on Symmetric Multiproeessor systems(SMP),based upon an antidiagonal motion approach for overcoming the data dependencies.The proposed algorithm is implemented using OpenMP directives.Numerical results, such as speedups and efficiency, are presented illustrating the efficient performance on a symmetric multiprocessor computer system, where the proposed algorithmic solution method achieves good speedups.
作者 张哲
出处 《计算机工程与应用》 CSCD 北大核心 2010年第29期47-49,52,共4页 Computer Engineering and Applications
关键词 稀疏线性系统 正规化精确并行逆算法 OPENMP sparse linear system parallel normalized exact inverse algorithms OpenMP
  • 相关文献

参考文献9

  • 1Lipitakis E A.Anormalized sparse linear equation solver[J].J Comput Appl Math, 1983,9:287-298.
  • 2Lipitakis E A, Evans D J.Solving linear finite element systems by normalized approximate matrix factorization semi-direct methods[J]. Comput Methods Appl Mech Eng, 1984,43 : 1-19.
  • 3Grawanis G A, Epitropou V N.On the performance of parallel approximate inverse preconditioning using Java multithreading techniques[J].Appl Math Comput, 2007, 190: 255-270.
  • 4Giannoutakis K M.Parallel exact and approximate arrow-type inverses on symmetric multiprocessor systems[C]//LNCS 3991: ICCS 2006,Part 1,2006.
  • 5Giavvanis G A,Giannoutakis K M.Normalized finite clement approximate inverse preconditioning for solving non-linear boundary value problems[M]//Bathc K J.Computational Fluid and Solid Mechanics 2003.Amsterdam :Elsevier,2003 : 1963-1967.
  • 6Huckle T.Approximate sparsity patterns for the inverse of a matrix and preconditioning[J].Appl Numer Math, 1999,30: 291-303.
  • 7Huclde T.Efficient computations of sparse approximate inverses[J]. Numer Linear Algebra Appl, 1998,5:57-71.
  • 8Akl S G.Parallel computation: Models and methods[M].Englewood Cliffs:Prentice Hall, 1997.
  • 9Dongarra J J, Duff I, Sorensen D.Numerical linear algebra for high-performance computers[M].Philadelphia:SIAM, 1998.

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部