Communication and Topology-aware Load Balancing in Charm++ with TreeMatch
JEANNOT, Emmanuel
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
MERCIER, Guillaume
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Voir plus >
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
JEANNOT, Emmanuel
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
MERCIER, Guillaume
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
TESSIER, François
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
< Réduire
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Langue
en
Communication dans un congrès
Ce document a été publié dans
IEEE Cluster 2013, 2013-09-24, Indianapolis. 2013-09-24
IEEE
Résumé en anglais
Programming multicore or manycore architectures is a hard challenge particularly if one wants to fully take advantage of their computing power. Moreover, a hierarchical topology implies that communication performance is ...Lire la suite >
Programming multicore or manycore architectures is a hard challenge particularly if one wants to fully take advantage of their computing power. Moreover, a hierarchical topology implies that communication performance is heterogeneous and this characteristic should also be exploited. We developed two load balancers for Charm++ that take into account both aspects, depending on the fact that the application is compute-bound or communication-bound. This work is based on our TREEMATCH library that computes process placement in order to reduce an application communication costs based on the hardware topology. We show that the proposed load-balancing schemes manage to improve the execution times for the two aforementioned classes of parallel applications.< Réduire
Mots clés en anglais
process placement
load balancing
message passing
communication cost
Origine
Importé de halUnités de recherche