Harnessing clusters of hybrid nodes with a sequential task-based programming model

Abstract : Runtime systems usually abstract a single node. The Sequential Task Flow (STF) model has been proven efficient on shared memory applications. When harnessing cluster of nodes, how should they communicate? By using explicit MPI user calls ? By using a specific paradigm ? Or can we keep the same STF paradigm and almost the same code, and leave the runtime system handle data transfers? We show how such a system has been sucessfully implemented on top of the StarPU runtime.
Type de document :
Communication dans un congrès
International Workshop on Parallel Matrix Algorithms and Applications (PMAA 2014), Jul 2014, Lugano, Switzerland. 2014
Liste complète des métadonnées

https://hal.inria.fr/hal-01283949
Contributeur : Marc Sergent <>
Soumis le : lundi 7 mars 2016 - 10:35:13
Dernière modification le : mardi 13 décembre 2016 - 15:40:50
Document(s) archivé(s) le : dimanche 13 novembre 2016 - 08:10:42

Fichier

pmaa14.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01283949, version 1

Citation

Emmanuel Agullo, Olivier Aumage, Mathieu Faverge, Nathalie Furmento, Florent Pruvost, et al.. Harnessing clusters of hybrid nodes with a sequential task-based programming model. International Workshop on Parallel Matrix Algorithms and Applications (PMAA 2014), Jul 2014, Lugano, Switzerland. 2014. <hal-01283949>

Partager

Métriques

Consultations de
la notice

260

Téléchargements du document

101