Efficient implementation of Jacobi iterative method for large sparse linear systems on graphic processing units

Abstract : In this paper, an original Jacobi implementation is considered for the solution of sparse linear systems of equations. The proposed algorithm helps to optimize the parallel implementation on GPU. The performance analysis of GPU-based (using CUDA) algorithm of the implementation of this algorithm is compared to the corresponding serial CPU-based algorithm. Numerical experiments performed on a set of matrices arising from the finite element discretization of various equations (3D Laplace equation, 3D gravitational potential equation, 3D Heat equation) with different meshes, illustrate the performance, robustness and efficiency of our algorithm, with a speed up to 23\times in double-precision arithmetics.
Type de document :
Article dans une revue
Journal of Supercomputing, Springer Verlag, 2017, 73 (8), pp.3411-3432. 〈10.1007/s11227-016-1701-3〉
Liste complète des métadonnées

https://hal-centralesupelec.archives-ouvertes.fr/hal-01708749
Contributeur : Frédéric Magoulès <>
Soumis le : mercredi 14 février 2018 - 09:36:13
Dernière modification le : lundi 14 mai 2018 - 07:41:18

Identifiants

Citation

Abal-Kassim Cheik Ahamed, Frédéric Magoulès. Efficient implementation of Jacobi iterative method for large sparse linear systems on graphic processing units. Journal of Supercomputing, Springer Verlag, 2017, 73 (8), pp.3411-3432. 〈10.1007/s11227-016-1701-3〉. 〈hal-01708749〉

Partager

Métriques

Consultations de la notice

31