Matrix Analysis and Parallel ComputingMakoto Natori, Takashi Nodera |
Contents
Preface | 1 |
Efficient Implementation of Multigrid Preconditioned Conjugate Gradient | 13 |
Parallel Computation for Parametric Study | 14 |
20 other sections not shown
Common terms and phrases
algorithm Analysis and Parallel applied approximation banyan network Bi-CG Bi-CGSTAB bound broadcasting calculation CG method coefficient matrix columns composite step condition number conjugate gradient method crossbar switch data transfer decomposition defined denote diagonal discretization eddy current edge element eigenvalues evaluate Figure filter factors finite element finite precision Gauss-Seidel method Gaussian Elimination ICCG ill-posed problems implementation iterative methods iterative solvers Jacobi Jacobi method Keio University Krylov subspace latency linear equations linear systems loop memory modules MFlops MGCG method multigrid method network topologies nonsymmetric number of iterations number of processors operations orthogonal packets Parallel Computing parallel prefix parameter PBSF performance polynomial preconditioned preconditioner problem PU's relative error Ritz values serial shared memory SIAM singular values solution solving sparse sparse matrix speedup SSS-MIN subspace symmetric TBSF technique Theorem tridiagonal tridiagonal matrix update variable workstation