[en] 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).
Disciplines :
Sciences informatiques
Auteur, co-auteur :
HELBECQUE, Guillaume ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > PCOG ; Université de Lille, CNRS/CRIStAL UMR 9189, Centre Inria de l’Université de Lille, France
GMYS, Jan; Université de Lille, CNRS/CRIStAL UMR 9189, Centre Inria de l’Université de Lille, France
CARNEIRO PESSOA, Tiago ; University of Luxembourg > Faculty of Science, Technology and Medicine > Department of Computer Science > Team Pascal BOUVRY
MELAB, Nouredine; Université de Lille, CNRS/CRIStAL UMR 9189, Centre Inria de l’Université de Lille, France
BOUVRY, Pascal ; University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Computer Science (DCS)
Co-auteurs externes :
yes
Langue du document :
Anglais
Titre :
Towards a Scalable Load Balancing for Productivity-Aware Tree-Search
Date de publication/diffusion :
02 juin 2023
Nom de la manifestation :
The 10th Annual Chapel Implementers and Users Workshop
