Interference-Aware Scheduling using Geometric Constraints - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année :

Interference-Aware Scheduling using Geometric Constraints

(1) , (2) , (2) , (2) , (2)
1
2

Résumé

The large scale parallel and distributed platforms produce a continuously increasing amount of data which have to be stored, exchanged and used by various jobs allocated on dierent nodes of the platform. The management of this huge communication demand is crucial for the performance of the system. Meanwhile, we have to deal with more interferences as the trend is to use a single all-purpose intercon-nection network. In this paper, we consider two dierent types of communications: the ows induced by data exchanges during computations and the ows related to Input/Output operations. We propose a general model for interference-aware scheduling, where explicit communications are replaced by external topological constraints. Specically, we limit the interferences of both communication types by adding geometric constraints on the allocation of jobs into machines. The proposed constraints reduce implicitly the data movements by restricting the set of possible allocations for each job. We present this methodology on the case study of simple network topologies, namely the line and the ring. We propose theoretical lower and upper bounds under dierent assumptions with respect to the platform and jobs characteristics. The obtained results illustrate well the diculty of the problem even on simple topologies.
Fichier principal
Vignette du fichier
convex-llncs.pdf (291.39 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-01884542 , version 1 (01-10-2018)

Identifiants

Citer

Raphaël Bleuse, Konstantinos Dogeas, Giorgio Lucarelli, Grégory Mounié, Denis Trystram. Interference-Aware Scheduling using Geometric Constraints. Euro-Par 2018 - European Conference on Parallel Processing, Aug 2018, Torino, Italy. pp.205-217, ⟨10.1007/978-3-319-96983-1_15⟩. ⟨hal-01884542⟩
234 Consultations
200 Téléchargements

Altmetric

Partager

Gmail Facebook Twitter LinkedIn More