Volume 5, Issue 1
Accelerating Preconditioned Iterative Linear Solvers on GPU

Hui Liu, Zhangxin Chen & Bo Yang

DOI:

Int. J. Numer. Anal. Mod. B, 5 (2014), pp. 136-146

Published online: 2014-05

Preview Full PDF 347 1054
Export citation
  • Abstract

Linear systems are required to solve in many scientific applications and the solution of these systems often dominates the total running time. In this paper, we introduce our work on developing parallel linear solvers and preconditioners for solving large sparse linear systems using NVIDIA GPUs. We develop a new sparse matrix-vector multiplication kernel and a sparse BLAS library for GPUs. Based on the BLAS library, several Krylov subspace linear solvers, and algebraic multi-grid (AMG) solvers and commonly used preconditioners are developed, including GMRES, CG, BICGSTAB, ORTHOMIN, classical AMG solver, polynomial preconditioner, ILU(k) and ILUT preconditioner, and domain decomposition preconditioner. Numerical experiments show that these linear solvers and preconditioners are efficient for solving the large linear systems.

  • Keywords

  • AMS Subject Headings

  • Copyright

COPYRIGHT: © Global Science Press

  • Email address
  • BibTex
  • RIS
  • TXT
@Article{IJNAMB-5-136, author = {Hui Liu, Zhangxin Chen and Bo Yang}, title = { Accelerating Preconditioned Iterative Linear Solvers on GPU}, journal = {International Journal of Numerical Analysis Modeling Series B}, year = {2014}, volume = {5}, number = {1}, pages = {136--146}, abstract = {Linear systems are required to solve in many scientific applications and the solution of these systems often dominates the total running time. In this paper, we introduce our work on developing parallel linear solvers and preconditioners for solving large sparse linear systems using NVIDIA GPUs. We develop a new sparse matrix-vector multiplication kernel and a sparse BLAS library for GPUs. Based on the BLAS library, several Krylov subspace linear solvers, and algebraic multi-grid (AMG) solvers and commonly used preconditioners are developed, including GMRES, CG, BICGSTAB, ORTHOMIN, classical AMG solver, polynomial preconditioner, ILU(k) and ILUT preconditioner, and domain decomposition preconditioner. Numerical experiments show that these linear solvers and preconditioners are efficient for solving the large linear systems.}, issn = {}, doi = {https://doi.org/}, url = {http://global-sci.org/intro/article_detail/ijnamb/225.html} }
TY - JOUR T1 - Accelerating Preconditioned Iterative Linear Solvers on GPU AU - Hui Liu, Zhangxin Chen & Bo Yang JO - International Journal of Numerical Analysis Modeling Series B VL - 1 SP - 136 EP - 146 PY - 2014 DA - 2014/05 SN - 5 DO - http://dor.org/ UR - https://global-sci.org/intro/article_detail/ijnamb/225.html KW - AB - Linear systems are required to solve in many scientific applications and the solution of these systems often dominates the total running time. In this paper, we introduce our work on developing parallel linear solvers and preconditioners for solving large sparse linear systems using NVIDIA GPUs. We develop a new sparse matrix-vector multiplication kernel and a sparse BLAS library for GPUs. Based on the BLAS library, several Krylov subspace linear solvers, and algebraic multi-grid (AMG) solvers and commonly used preconditioners are developed, including GMRES, CG, BICGSTAB, ORTHOMIN, classical AMG solver, polynomial preconditioner, ILU(k) and ILUT preconditioner, and domain decomposition preconditioner. Numerical experiments show that these linear solvers and preconditioners are efficient for solving the large linear systems.
Hui Liu, Zhangxin Chen & Bo Yang. (1970). Accelerating Preconditioned Iterative Linear Solvers on GPU. International Journal of Numerical Analysis Modeling Series B. 5 (1). 136-146. doi:
Copy to clipboard
The citation has been copied to your clipboard