A contention-aware performance model for HPC-based networks: A case study of the InfiniBand network

Maxime Martinasso 1 Jean-François Méhaut 1
1 MESCAL - Middleware efficiently scalable
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
Abstract : Multi-core clusters are cost-effective clusters largely used in high-performance computing. Parallel applications using message passing as a communication mechanism may introduce complex communication behaviours on such clusters. By sending and receiving data simultaneously from and to several nodes, parallel applications create concurrent accesses to the resources of the network. In this paper, we present a general model that expresses network resource sharing characterised by a dynamic contention graph. The model is based on a linear system weighted by bandwidth distribution factors called penalty coefficients that are specific to a network technology. We propose a method to solve the linear system and present an analysis to determine penalty coefficients on InfiniBand technology. We use complex network conflicts to assess the ability of the model to predict with low errors.
Type de document :
Communication dans un congrès
Springer. Euro-Par 2011 : Proceedings of the 17th International Euro-Par Conference, Aug 2011, Bordeaux, France. Springer, 6852, pp.91-102, 2011, LNCS. 〈10.1007/978-3-642-23400-2_10〉
Liste complète des métadonnées
https://hal.inria.fr/hal-00690876
Contributeur : Ist Rennes <>
Soumis le : mardi 24 avril 2012 - 15:56:10
Dernière modification le : jeudi 11 octobre 2018 - 08:48:02

Identifiants

Collections

INRIA | LIG | GRID5000 | UGA | LIG_SRCPR

Citation

Maxime Martinasso, Jean-François Méhaut. A contention-aware performance model for HPC-based networks: A case study of the InfiniBand network. Springer. Euro-Par 2011 : Proceedings of the 17th International Euro-Par Conference, Aug 2011, Bordeaux, France. Springer, 6852, pp.91-102, 2011, LNCS. 〈10.1007/978-3-642-23400-2_10〉. 〈hal-00690876〉

Exporter

BibTeX TEI DC DCterms EndNote

Partager

Métriques

Consultations de la notice

207

Contact


archive-ouverte@inria.fr