Parallel Two-Dimensional Quicksort Algorithm (PTSA)

Abstract

In this research, a Parallel Two-Dimensional Sorting Algorithm (PTSA) is presented that has better performance than the classical Quicksort, to sort a data vector of size n = r (rows) × c(columns). PTSA algorithm divides the input vector into n/r sub-vectors, which represents tow-dimensional vector of Slave Processor Elements (VPE ), the maximum number of VPE for parallel sorting is equal to r×c, VPE just the read,and write operations. The number of Master Processors (MP) which do the sort operation is equal to c, The time needed for PTSA algorithm is reduced by θ (n/r log n/r) with respect to the time needed by Quicksort θ (n log n) to sort the same vector. Simulation results show that the efficiency of sorting using PTSA algorithm is increased and the complexity is reduced significantly compared with classical Quicksort.