An Experimental Performance Study of Polynomial Preconditioner in PCG
No Thumbnail Available
Date
2022
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
none
Solving laplacian system is common in the field of computer science nowadays. Preconditioner is an essential tool while solving linear system with indirect method. It may bring significant improvements to number of iterations, CPU time, and the errors. In this work we will start with graph laplacian, matrix splitting and approximation theories to get some polynomial preconditioners, and investigate the performance in the changes of different parameters in PCG(Preconditioned Conjugate Gradient) method mainly by experiments. We will show the experimental result as conclusion for the purpose of accelerating the iteration in future works.
Solving laplacian system is common in the field of computer science nowadays. Preconditioner is an essential tool while solving linear system with indirect method. It may bring significant improvements to number of iterations, CPU time, and the errors. In this work we will start with graph laplacian, matrix splitting and approximation theories to get some polynomial preconditioners, and investigate the performance in the changes of different parameters in PCG(Preconditioned Conjugate Gradient) method mainly by experiments. We will show the experimental result as conclusion for the purpose of accelerating the iteration in future works.
Description
Keywords
none, Graph Laplacian, PCG, Polynomial Preconditioner