Co-scheduling Amdahl applications on cache-partitioned systems

Guillaume Aupy 1 Anne Benoit 2, 3 Sicheng Dai 4 Loïc Pottier 2, 3 Padma Raghavan 5 Yves Robert 2, 3 Manu Shantharam 6
1 TADAAM - Topology-Aware System-Scale Data Management for High-Performance Computing
LaBRI - Laboratoire Bordelais de Recherche en Informatique, Inria Bordeaux - Sud-Ouest
2 LIP - Laboratoire de l'Informatique du Parallélisme
3 ROMA - Optimisation des ressources : modèles, algorithmes et ordonnancement
Inria Grenoble - Rhône-Alpes, LIP - Laboratoire de l'Informatique du Parallélisme
4 ECNU - East China Normal University [Shangaï]
5 Vanderbilt University [Nashville]
6 San Diego Supercomputer Center [San Diego]
Abstract : Cache-partitioned architectures allow subsections of the shared last-level cache (LLC) to be exclusively reserved for some applications. This technique dramatically limits interactions between applications that are concurrently executing on a multi-core machine. Consider n applications that execute concurrently, with the objective to minimize the makespan, defined as the maximum completion time of the n applications. Key scheduling questions are: (i) which proportion of cache and (ii) how many processors should be given to each application? In this paper, we provide answers to (i) and (ii) for Amdahl applications. Even though the problem is shown to be NP-complete, we give key elements to determine the subset of applications that should share the LLC (while remaining ones only use their smaller private cache). Building upon these results, we design efficient heuristics for Amdahl applications. Extensive simulations demonstrate the usefulness of co-scheduling when our efficient cache partitioning strategies are deployed.
Keywords : complexity results cache partitioning Co-scheduling
Type de document :
Article dans une revue
International Journal of High Performance Computing Applications, SAGE Publications, 2017, 〈10.1177/1094342017710806〉
Liste complète des métadonnées

Littérature citée [25 références]  Voir  Masquer  Télécharger
https://hal.archives-ouvertes.fr/hal-01670137
Contributeur : Equipe Roma <>
Soumis le : jeudi 21 décembre 2017 - 10:26:05
Dernière modification le : vendredi 20 avril 2018 - 15:44:27

Fichier

ijhpca-revised.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

UNIV-BORDEAUX | LABRI | LIP | ENS-LYON | INRIA | INRIA2017 | UNIV-LYON1

Citation

Guillaume Aupy, Anne Benoit, Sicheng Dai, Loïc Pottier, Padma Raghavan, et al.. Co-scheduling Amdahl applications on cache-partitioned systems. International Journal of High Performance Computing Applications, SAGE Publications, 2017, 〈10.1177/1094342017710806〉. 〈hal-01670137〉

Exporter

BibTeX TEI DC DCterms EndNote

Partager

Métriques

Consultations de la notice

264

Téléchargements de fichiers

74

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD