Accéder directement au contenu Accéder directement à la navigation
Article dans une revue

Parallelizing with BDSC, a resource-constrained scheduling algorithm for shared and distributed memory systems

Abstract : We introduce a new parallelization framework for scientific computing based on BDSC, an efficient automatic scheduling algorithm for parallel programs in the presence of resource constraints on the number of processors and their local memory size. BDSC extends Yang and Gerasoulis's Dominant Sequence Clus-tering (DSC) algorithm; it uses sophisticated cost models and addresses both shared and distributed parallel memory architectures. We describe BDSC, its integration within the PIPS compiler infrastructure and its application to the parallelization of four well-known scientific applications: Harris, ABF, equake and IS. Our experiments suggest that BDSC's focus on efficient resource man-agement leads to significant parallelization speedups on both shared and dis-tributed memory systems, improving upon DSC results, as shown by the com-parison of the sequential and parallelized versions of these four applications running on both OpenMP and MPI frameworks.
Liste complète des métadonnées

Littérature citée [31 références]  Voir  Masquer  Télécharger

https://hal-mines-paristech.archives-ouvertes.fr/hal-01097328
Contributeur : Claire Medrala <>
Soumis le : vendredi 19 décembre 2014 - 14:22:07
Dernière modification le : jeudi 24 septembre 2020 - 16:36:02
Archivage à long terme le : : lundi 23 mars 2015 - 17:52:09

Fichier

A-499-accepte.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Dounia Khaldi, Pierre Jouvelot, Corinne Ancourt. Parallelizing with BDSC, a resource-constrained scheduling algorithm for shared and distributed memory systems. Parallel Computing, Elsevier, 2015, 41, pp.66 - 89. ⟨10.1016/j.parco.2014.11.004⟩. ⟨hal-01097328⟩

Partager

Métriques

Consultations de la notice

286

Téléchargements de fichiers

383