Impact study of data locality on task-based applications through the Heteroprio scheduler - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Article Dans Une Revue PeerJ Computer Science Année : 2019

Impact study of data locality on task-based applications through the Heteroprio scheduler

Résumé

The task-based approach has emerged as a viable way to effectively use modern heterogeneous computing nodes. It allows the development of parallel applications with an abstraction of the hardware by delegating task distribution and load balancing to a dynamic scheduler. In this organization, the scheduler is the most critical component that solves the DAG scheduling problem in order to select the right processing unit for the computation of each task. In this work, we extend our Heteroprio scheduler that was originally created to execute the fast multipole method on multi-GPUs nodes. We improve Heteroprio by taking into account data locality during task distribution. The main principle is to use different task-lists for the different memory nodes and to investigate how locality affinity between the tasks and the different memory nodes can be evaluated without looking at the tasks' dependencies. We evaluate the benefit of our method on two linear algebra applications and a stencil code. We show that simple heuristics can provide significant performance improvement and cut by more than half the total memory transfer of an execution.
Fichier principal
Vignette du fichier
peerj-cs-190.pdf (2.47 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02120736 , version 1 (06-05-2019)

Identifiants

Citer

Bérenger Bramas. Impact study of data locality on task-based applications through the Heteroprio scheduler. PeerJ Computer Science, 2019, 5, pp.e190. ⟨10.7717/peerj-cs.190⟩. ⟨hal-02120736⟩
141 Consultations
182 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More