LU Factorization for Accelerator-based Systems

Abstract : Multicore architectures enhanced with multiple GPUs are likely to become mainstream High Performance Computing (HPC) platforms in a near future. In this paper, we present the design and implementation of an LU factorization using tile algorithm that can fully exploit the potential of such platforms in spite of their complexity. We use a methodology derived from previous work on Cholesky and QR factor- izations. Our contributions essentially consist of providing new CPU/GPU hybrid LU kernels, studying the impact on performance of the looking variants as well as the storage layout in presence of pivoting, tuning the kernels for two different machines composed of multiple recent NVIDIA Tesla S1070 (four GPUs total) and Fermi-based S2050 GPUs (three GPUs total), respectively. The hybrid tile LU asymptotically achieves 1 Tflop/s in single precision on both hardwares. The performance in double precision arithmetic reaches 500 Gflop/s on the Fermi-based system, twice faster than the old GPU generation of Tesla S1070. We also discuss the impact of the number of tiles on the numerical stability. We show that the numerical results of the tile LU factorization will be accurate enough for most applications as long as the computations are performed in double precision arithmetic.
Type de document :
Communication dans un congrès
9th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA 11), Jun 2011, Sharm El-Sheikh, Egypt. 2011
Liste complète des métadonnées


https://hal.inria.fr/hal-00654193
Contributeur : Samuel Thibault <>
Soumis le : mercredi 21 décembre 2011 - 10:24:09
Dernière modification le : jeudi 10 septembre 2015 - 01:08:11
Document(s) archivé(s) le : vendredi 16 novembre 2012 - 16:16:12

Fichier

plagma_lu.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00654193, version 1

Collections

Citation

Emmanuel Agullo, Cédric Augonnet, Jack Dongarra, Mathieu Faverge, Julien Langou, et al.. LU Factorization for Accelerator-based Systems. 9th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA 11), Jun 2011, Sharm El-Sheikh, Egypt. 2011. <hal-00654193>

Partager

Métriques

Consultations de
la notice

727

Téléchargements du document

363