Co-scheduling HPC workloads on cache-partitioned CMP platforms

Abstract : Co-scheduling techniques are used to improve the throughput of applications on chip multiprocessors (CMP), but sharing resources often generates critical interferences. We focus on the interferences in the last level of cache (LLC) and use the Cache Allocation Technology (CAT) recently provided by Intel to partition the LLC and give each co-scheduled application their own cache area. We consider m iterative HPC applications running concurrently and answer the following questions: (i) how to precisely model the behavior of these applications on the cache partitioned platform? and (ii) how many cores and cache fractions should be assigned to each application to maximize the platform efficiency? Here, platform efficiency is defined as maximizing the performance either globally, or as guaranteeing a fixed ratio of iterations per second for each application. Through extensive experiments using CAT, we demonstrate the impact of cache partitioning when multiple HPC application are co-scheduled onto CMP platforms.
Type de document :
Communication dans un congrès
IEEE Cluster 2018, Sep 2018, Belfast, United Kingdom. pp.335-345, 2018, Proceedings the 20th IEEE Cluster Conference
Liste complète des métadonnées

Littérature citée [9 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01874154
Contributeur : Brice Goglin <>
Soumis le : vendredi 14 septembre 2018 - 09:56:11
Dernière modification le : vendredi 21 septembre 2018 - 17:10:27

Fichier

cluster18.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01874154, version 1

Citation

Guillaume Aupy, Anne Benoit, Brice Goglin, Loïc Pottier, Yves Robert. Co-scheduling HPC workloads on cache-partitioned CMP platforms. IEEE Cluster 2018, Sep 2018, Belfast, United Kingdom. pp.335-345, 2018, Proceedings the 20th IEEE Cluster Conference. 〈hal-01874154〉

Partager

Métriques

Consultations de la notice

107

Téléchargements de fichiers

19