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).