Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Computational Science, Engineering & Technology Series
PATTERNS FOR PARALLEL PROGRAMMING ON GPUS
Edited by: F. Magoulès
Parallel Preconditioned Conjugate Gradient Algorithm on GPU
F. Andzembe and J. Koko
LIMOS, Université Blaise Pascal – CNRS UMR 6158, Clermont-Ferrand, France
F. Andzembe, J. Koko, "Parallel Preconditioned Conjugate Gradient Algorithm on GPU", in F. Magoulès, (Editor), "Patterns for Parallel Programming on GPUs", Saxe-Coburg Publications, Stirlingshire, UK, Chapter 9, pp 209-225, 2014. doi:10.4203/csets.34.9
Keywords: preconditioned conjugate gradient, parallel computing, graphics processing unit.
We are proposing a parallel implementation of the preconditioned conjugate gradient algorithm on a GPU-platform. The preconditioning matrix is a first order approximate inverse derived from the SSOR preconditioner. Used through sparse matrix-vector multiplication, the proposed preconditioner is well-suited for massively parallel architectures like GPUs. Compared to CPU implementation of the conjugate gradient algorithm, our GPU preconditioned conjugate gradient implementation is up to 16 times faster (8 times faster at worst).
purchase the full-text of this chapter (price £20)