Open Access Open Access  Restricted Access Subscription or Fee Access

A Survey on Optimization and Parallelization of Conjugate Gradient Solver

Puja P. Khirodkar, Vikas Kumar, P. P. Joshi

Abstract


-Conjugate Gradient (CG) Solver is a well-known iterative techniques for solving sparse symmetric positive definite (SPD) systems of linear equations. The aim of this paper is to introduce rarely used technique to optimize and parallelize the currently available Conjugate Gradient Solver on GPU using CUDA which stands for Compute Unified Device Architecture. Existing Conjugate Gradient Solver can be optimized with the help of some techniques available for sparse matrix storage like Compressed Sparse Vector (CSV). CUDA is a parallel computing platform and application programming interface (API) model created by NVIDIA. Threads can be used for parallel execution of the iterative part of CG solver. This parallelization will definitely speedup the performance of CG solver which can be used in many compute fluid dynamic (CFD) computations.


Keywords


Iterative Methods, Convergence, Sparse and Very Large Systems, Linear Systems, Parallel Programming.

Full Text:

PDF

References


Torsten Hoeer,”Optimizing a conjugate gradient solver with non-blocking collective operations”, PVM/MPI 2006, LNCS 4192, pp 374-382 2006.

Tom R. Halfhill ,”Parallel processing with CUDA:WhitePaper”

Pierre Kestener, ”Introduction to massively parallel computing with GPU:Whitepaper”

”Whitepaper: NVIDIAs Next generation CUDA compute architecture: Fermi”

Ana Flvia Peixoto de Camargos, Viviane Cristine Silva, Jean-Michel Guichon, and Grard Munier, ”Efficient parallel preconditioned conjugate gradient solver on GPU for FE modeling of electromagnetic fields highly dissipative media” , IEEE TRANSACTIONS ON MAGNETICS, VOL. 50, NO. 2, FEBRUARY 2014 7014004.

Yao Chen, Yonghua Zhao, Wei Zhao, Lian Zhao ”A Comparative Study of preconditioners for GPU-accelerated conjugate gradient solver”, 2013 IEEE International Conference on High Performance Computing and Communications 2013 IEEE International Conference on Embedded and Ubiquitous Computing .

A.Wendell de O. Rodrigues, Loc Chevallier,”Test harness on a preconditioned conjugate gradient solver on GPUs:An Efficiency Analysis”,IEEE TRANSACTIONS ON MAGNETICS, VOL. 49, NO.5, MAY 2013.

G. Oyarzun, R. Borrell,”MPI-CUDA sparse matrix vector multiplication for the conjugate” ,Computers Fluids 92 (2014) 244252.

Guiming Wu, Xianghui Xie, Member, IEEE, Yong Dou, Member, IEEE, and Miao Wang,”High-performance architecture for the conjugate gradient solver on FPGAs”, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMSII: EXPRESS BRIEFS, VOL. 60, NO. 11, NOVEMBER 2013.

Maryam Mehri Dehnavi, David M. Fernández, and Dennis Giannacopoulos,”Enhancing the Performance of Conjugate Gradient Solvers on Graphic Processing Units”, IEEE TRANSACTIONS ON MAGNETICS, VOL. 47, NO. 5, MAY 2011.

Dominik Michels,”Sparse-Matrix-CG-Solver in CUDA”, Proceedings of CESCG 2011: The 15th Central European Seminar on Computer Graphics.

Ament, G.Knittel, D.Weiskopf, W.Straber,”A parallel preconditioned conjugate gradient solver for the poisson problem on a multi-gpu platform”, Parallel, Distributed and Network-Based Processing (PDP), 2010 18th Euromicro International Conference on. IEEE, 2010.

Aiyoub Farzaneh, Hossein Kheiri and Mehdi Abbaspour Shahmersi,”An efficient storage format for large sparse matrices”, Commun. Fac. Sci . Univ. Ank. Series A1 Volume 58, Number 2, Pages 110 (2009) ISSN 13035991.

N. Bell and M. Garland,”Efficient sparse matrix-vector multiplication on CUDA”, NVIDIA Corporation, Santa Clara, CA, USA, NVIDIA Tech. Rep. NVR-2008-004, Dec. 2008.

Y. Saad,”Iterative methods for sparse linear systems”, 2nd ed. New York,NY, USA: PWS, 2003, pp. 275369.


Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.