Towards Reproducible Blocked LU Factorization - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

Towards Reproducible Blocked LU Factorization

Résumé

In this article, we address the problem of reproducibility of the blocked LU factorization on GPUs due to cancellations and rounding errors when dealing with floating-point arithmetic. Thanks to the hierarchical structure of linear algebra libraries, the computations carried within this operation can be expressed in terms of the Level-3 BLAS routines as well as the unblocked variant; the latter is correspon-dently built upon the Level-1/2 BLAS kernels. In addition, we strengthen numerical stability of the blocked LU factorization via partial row pivoting. Therefore, we propose a double-layer bottom-up approach for ensuring reproducibility of the blocked LU factorization and provide experimental results for its underlying blocks.
Fichier principal
Vignette du fichier
REPPAR-05.pdf (311.92 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01456307 , version 1 (04-02-2017)
hal-01456307 , version 2 (22-03-2017)

Identifiants

  • HAL Id : hal-01456307 , version 2

Citer

Roman Iakymchuk, Enrique S Quintana-Ortí, Erwin S Laure, Stef S Graillat. Towards Reproducible Blocked LU Factorization. 4th International Workshop on Reproducibility in Parallel Computing in conjunction with IPDPS 2017 - 31st IEEE International Parallel & Distributed Processing Symposium, May 2017, Orlando, United States. ⟨hal-01456307v2⟩
211 Consultations
360 Téléchargements

Partager

Gmail Facebook X LinkedIn More