Order Statistics on a Hypercube

We consider an order statistics problem, a variant of the selection problem. We present an optimal parallel algorithm to find the highest k elements out of a set of n totally ordered (but not sorted) elements. This algorithm runs on a hypercube multiprocessor in O(n/p) time (p is the number of processors) using up to O(n1-θ) processors, 0<θ<1. We assume that the overriding communication cost is in the setting up of the message transfer between nodes.

[1]  Richard Cole,et al.  A Parallel Median Algorithm , 1985, Inf. Process. Lett..

[2]  János Komlós,et al.  Deterministic selection in O(loglog N) parallel time , 1986, STOC '86.

[3]  Charles L. Seitz,et al.  The cosmic cube , 1985, CACM.

[4]  Kurt Mehlhorn,et al.  Deterministic Simulation of Idealized Parallel Computers on More Realistic Ones , 1986, SIAM J. Comput..

[5]  Harold S. Stone,et al.  Parallel Processing with the Perfect Shuffle , 1971, IEEE Transactions on Computers.

[6]  Franco P. Preparata,et al.  The cube-connected-cycles: A versatile network for parallel computation , 1979, 20th Annual Symposium on Foundations of Computer Science (sfcs 1979).

[7]  Larry S. Davis,et al.  The Hough Transform on the Butterfly and the NCUBE. , 1986 .

[8]  Michael J. Flynn,et al.  Some Computer Organizations and Their Effectiveness , 1972, IEEE Transactions on Computers.

[9]  Leslie G. Valiant,et al.  Parallelism in Comparison Problems , 1975, SIAM J. Comput..

[10]  Selim G. Akl.  An Optimal Algorithm for Parallel Selection , 1984, Inf. Process. Lett..

[11]  Marshall C. Pease,et al.  The Indirect Binary n-Cube Microprocessor Array , 1977, IEEE Transactions on Computers.