Load balanced priority queues on distributed memory machines

Abstract

We consider efficient algorithms for priority queues on distributed memory multiprocessors, such as nCUBE, iPSc, MPP and loosely-coupled systems consisting of networked workstations. For a p-processor distributed memory multicomputer P and n data items in the priority queue, n>p. we investigate two priority queues; horizontally sliced and vertically sliced. Both of these achieve load balance, i.e. at most θ(n/p) data items are stored at every processor of P. Horizontally sliced priority queue allows deletions and insertions of θ(p) items in time O(p/bwΤ _c +Τ _p pp log n) on hypercubic networks where Τ _c is the communication time between a pair of processors, Τ _p is the unit processing time and bw is the width of the communication channel between a pair of processors. Vertically sliced priority queue allows deletions and insertions of θ(p) items in time O((Τ _c +Τ _p ) log p log n) on hypercubic networks. Similar results hold for other types of networks.