Enabling High-Performance Memory Migration for Multithreaded Applications on Linux

Brice Goglin 1, 2 Nathalie Furmento 1, 2
2 RUNTIME - Efficient runtime systems for parallel architectures
Inria Bordeaux - Sud-Ouest, UB - Université de Bordeaux, CNRS - Centre National de la Recherche Scientifique : UMR5800
Abstract : As the number of cores per machine increases, memory architectures are being redesigned to avoid bus contention and sustain higher throughput needs. The emergence of Non-Uniform Memory Access (NUMA) constraints has caused affinities between threads and buffers to become an important decision criteria for schedulers. Memory migration enables the dynamically joined distribution of work and data across the machine but requires high-performance data transfers as well as a convenient programming interface. We present the improvement of the Linux migration primitives and the implementation of a Next-Touch policy in the kernel to provide multithreaded applications with an easy way to dynamically maintain thread-data affinity. Microbenchmarks show that our work enables a high-performance, synchronous and lazy memory migration within multithreaded applications. A threaded LU factorization then reveals the large improvement that our Next-Touch policy model may bring in applications with complex access patterns.
Type de document :
Communication dans un congrès
IEEE. MTAAP'09: Workshop on Multithreaded Architectures and Applications, held in conjunction with IPDPS 2009, May 2009, Rome, Italy. 2009, <10.1109/IPDPS.2009.5161101>
Liste complète des métadonnées

https://hal.inria.fr/inria-00358172
Contributeur : Brice Goglin <>
Soumis le : mardi 3 février 2009 - 07:37:40
Dernière modification le : jeudi 10 septembre 2015 - 01:06:31
Document(s) archivé(s) le : vendredi 12 octobre 2012 - 10:30:51

Fichier

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

Identifiants

Collections

Citation

Brice Goglin, Nathalie Furmento. Enabling High-Performance Memory Migration for Multithreaded Applications on Linux. IEEE. MTAAP'09: Workshop on Multithreaded Architectures and Applications, held in conjunction with IPDPS 2009, May 2009, Rome, Italy. 2009, <10.1109/IPDPS.2009.5161101>. <inria-00358172>

Partager

Métriques

Consultations de
la notice

293

Téléchargements du document

621