SPAWN: An Iterative, Potentials-Based, Dynamic Scheduling and Partitioning Tool

Abstract : Many applications of physics modeling use regular meshes on which computations of highly variable cost can occur. Distributing the underlying cells over manycore architec-tures is a critical load balancing step that should increase the period until another step is required. Graph partitioning tools are known to be very effective for such problems, but they exhibit scalability problems as the number of cores and the number of cells increases. We introduce a dynamic task scheduling approach inspired by physical particles interactions. Our method allows cores to virtually move over a 2D/3D mesh of tasks and uses a Voronoi domain decomposition to balance workload among cores. Displacements of cores are the result of force computations using a carefully chosen pair potential. We evaluate our method against graph partitioning tools and existing task schedulers with a representative physical application, and demonstrate the relevance of our approach.
Type de document :
Communication dans un congrès
SuperComputing'15 - RESPA Workshop, Nov 2015, Austin, United States. 〈http://respa15.rice.edu/〉
Liste complète des métadonnées

Littérature citée [20 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01223897
Contributeur : Jean-Charles Papin <>
Soumis le : mardi 3 novembre 2015 - 16:11:10
Dernière modification le : jeudi 29 mars 2018 - 09:04:04
Document(s) archivé(s) le : jeudi 4 février 2016 - 11:21:46

Fichier

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

Identifiants

  • HAL Id : hal-01223897, version 1

Citation

Jean-Charles Papin, Christophe Denoual, Laurent Colombet, Raymond Namyst. SPAWN: An Iterative, Potentials-Based, Dynamic Scheduling and Partitioning Tool. SuperComputing'15 - RESPA Workshop, Nov 2015, Austin, United States. 〈http://respa15.rice.edu/〉. 〈hal-01223897〉

Partager

Métriques

Consultations de la notice

348

Téléchargements de fichiers

412