Affiliation:
1. National University of Defense Technology
2. China Institute of Water Resources and Hydropower Research
Abstract
The general parallel preconditioning subspace iteration software GPPS is developed based on the integration of various preconditioning methods, parallelization techniques, and subspace iterations, to solve general sparse linear systems. The software has several characteristics, including high computational performance, good readability, simple interface, and excellent scalability. In this paper, the functions of the software are outlined, and then it is used to solve the sparse linear systems from meso-scale simulation of concrete specimens. The numerical experiments show that GPPS is clearly superior to the software package AZTEC, which is used in the simulation up to now, and in addition, ICT and the parallelization technique based on factors combination outperform other preconditioning techniques integrated in the software.
Publisher
Trans Tech Publications, Ltd.
Reference8 articles.
1. Y. Saad. Iterative Methods for Sparse Linear Systems. PWS, Boston, (1996).
2. J. P. Wu, X. P. Liu, Z. H. Wang, Z. H. Dai, and X. M. Li, Two new preprocessing methods to 2-dimension energy equations with three temperatures, Chinese J. Comput. Phys., 2005, 22(4): 283–291.
3. Wu Jianping, Wang Zhenghua, Li Xiaomei. Problems and improvements to the incomplete Cholesky decomposition with thresholds. Journal on Numerical Methods and Computer Applications, 2003, 24(3): 207-214.
4. K. Nakajima, H. Okuda. Parallel Iterative Solvers with Localized ILU Preconditioning for Unstructured Grids on Workstation Clusters. International Journal of Computational Fluid Dynamics, 1999, 12(3): 315-322.
5. M. M. monga Made, H. A. van der Vorst. Parallel incomplete factorizations with pseudo-overlapped subdomains. Parallel Computing, 2001, 27: 989-1008.