Solving lattice QCD systems of equations using mixed precision solvers on GPUs
Harvard-Smithsonian Center for Astrophysics, 60 Garden St, Cambridge, MA 02138, USA
Computer Physics Communications, Volume 181, Issue 9, September 2010, Pages 1517-1528 (13 May 2010)
@article{clark2010solving,
title={Solving Lattice QCD systems of equations using mixed precision solvers on GPUs},
author={Clark, MA and Babich, R. and Barros, K. and Brower, RC and Rebbi, C.},
journal={Computer Physics Communications},
issn={0010-4655},
year={2010},
publisher={Elsevier}
}
Modern graphics hardware is designed for highly parallel numerical tasks and promises significant cost and performance benefits for many scientific applications. One such application is lattice quantum chromodynamics (lattice QCD), where the main computational challenge is to efficiently solve the discretized Dirac equation in the presence of an SU (3) gauge field. Using NVIDIA’s CUDA platform we have implemented a Wilson-Dirac sparse matrix-vector product that performs at up to 40, 135 and 212 Gflops for double, single and half precision respectively on NVIDIA’s GeForce GTX 280 GPU. We have developed a new mixed precision approach for Krylov solvers using reliable updates which allows for full double precision accuracy while using only single or half precision arithmetic for the bulk of the computation. The resulting BiCGstab and CG solvers run in excess of 100 Gflops and, in terms of iterations until convergence, perform better than the usual defect-correction approach for mixed precision.
November 8, 2010 by hgpu
