Selection-Based Parallel Range-Join in Hypercubes

In this paper we propose a new parallel algorithm for computing the range-join of two sets R and S, i.e., the set containing all tuples (r, s), r 2 R and s 2 S, such that e 1 jr ? sj e 2 , in a hypercube computer with p processors, where 0 e 1 e 2 are constants and 1 p jRj = m jSj = n. Our algorithm uses selection as a basis and works in divide-and-conquer fashion. The algorithm requires at most O(n) local memory and has a worst-case time complexity O(mn p). The best-case time complexity of the algorithm is O(m p log 2 p + n p log m) (or O(T med + n p log m) if a hypercube median-selection algorithm of time complexity T med is employed) which is better than permutation-based join developed recently 9] when n pm. An upper bound of time complexity in the general case and its full evaluation is also shown. The proposed algorithm works also for the case m < p n.