Keita Teranishi and Padma Raghavan
The Pennsylvania State University
USA
Abstract:
We report on the parallel performance of several preconditioning schemes to accelerate the convergence of Conjugate Gradients (CG) for sparse linear system solution. Our preconditioners are constructed and applied in parallel using incomplete factorizations with selective inversion for improved latency-tolerance. We provide empirical results on the efficiency, scalability and quality of our preconditioners for systems from nonlinear optimization and structural mechanics.