BPX-Like Preconditioned Conjugate Gradient Solvers for Poisson Problem and Their CUDA Implementations

Jie Peng,Shi Shu,Chunsheng Feng,Xiaoqiang Yue
DOI: https://doi.org/10.1007/978-3-319-38789-5_70
2017-01-01
Abstract:In this paper, we firstly introduce two BPX-like preconditioners B-J(1) and B-J(3), and present an equivalent but more robust BPX-like preconditioner B-J(2) for the solution of the linear finite element discretization of Poisson problem. Secondly, we implement these preconditioners and their preconditioned conjugate gradient (PCG) solvers B-l(p)-CG(l = 1, 2, 3) under Compute Unified Device Architecture (CUDA), where we exploit the hierarchical and the overall storage structure, take advantage of the multicolored Gauss-Seidel smoother. Finally, comparisons are made among these PCG solvers and the state-of-the-art SA-AMG preconditioned CG solver (SA-CG) in CUSP library. Numerical results demonstrate that the iteration numbers of B-2(p)-CG holds the weakest dependence on the grid size, while B-3(p)-CG is the most efficient solver. Furthermore, B-3(p)-CG possesses considerable advantages over SA-CG in computational capability and efficiency. In particular, B-3(p)-CG runs 3.67 times faster than SA-CG when solving a problem with about one-million unknowns.
What problem does this paper attempt to address?