Range search in parallel using distributed data structures

Abstract

We present a parallel algorithm to obtain a set of points in a rectangular parallelpiped (range-search) in O(log n) time, with only (2 log 2 n − 10 log n + 14) processors on a EREW-PRAM, where processors are allowed to communicate through messages. We also present a nontrivial implementation technique on the hypercube parallel architecture with which the above time and processor bounds can be achieved without any communication overhead. A parallel algorithm for range searching is developed here using the concept of distributed data structures. We use the range tree proposed by Bentley as our data structure to be distributed. Range search has important applications in the areas of databases and computational geometry.