Towards a scalable load balancing for productivity-aware tree-search - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Towards a scalable load balancing for productivity-aware tree-search

Résumé

In the context of exascale programming, we investigate a parallel distributed productivity-aware tree-search for exact optimization in Chapel. To this end, we present the DistBag-DFS distributed data structure, which is our revisited version of the Chapel’s DistBag data structure for depth-first search. The latter implements a distributed multi-pool, as well as an underlying locality-aware load balancing mechanism. Extensive experiments on large unbalanced tree-based problems are performed, and the competitiveness of our approach is reported against MPI+X implementations in terms of performance. For our best results, we achieve 94% of the ideal speed-up, using up to 64 computer nodes (8192 cores).
Fichier non déposé

Dates et versions

hal-04117159 , version 1 (05-06-2023)

Identifiants

  • HAL Id : hal-04117159 , version 1

Citer

Guillaume Helbecque, Jan Gmys, Tiago Carneiro, Nouredine Melab, Pascal Bouvry. Towards a scalable load balancing for productivity-aware tree-search. The 10th Annual Chapel Implementers and Users Workshop, Jun 2023, online, United States. ⟨hal-04117159⟩
27 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More