Design and Evaluation of Topology-aware Scatter and AllGather Algorithms for Dragonfly Networks
Résumé
High-radix direct network topologies such as Dragonfly have been proposed for petascale and exascale supercomputers because they ensure fast interconnections and reduce the cost of the network compared with traditional network topologies. However, current algorithms for communication do not consider the topology and thus lose numerous opportunities of optimization for performance. In this work, we exploit the strength of the Dragonfly with topology-aware algorithms for AllGather and Scatter operations. We study existing algorithms, then propose derived algorithms, that we evaluate using CODES, an event-driven simulator. As expected, for AllGather, making algorithms topology-aware does improve the performance and reduces the link utilization. However, simulations of various Scatter algorithms show surprising results, and point out the important role played by hardware for the efficiency of the algorithms. In particular, the knowledge of the number and size of input/output buffers in routers can be exploited to accelerate the Scatter operation by a factor up to 2 times.
Origine : Fichiers produits par l'(les) auteur(s)