A New Sorting Algorithm Based on Interpolation Merging and It’s Parallel Implementation

Abstract

We present a new modified merge sort algorithm (invented by John von Newmann in 1945) based on interpolation merging technique. Known to date, simple binary merging is the best general purpose merging algorithm. However, interpolation merging requires substantially fewer comparisons and increases the performance of modified merge sort algorithm. Results from a computer implementation of the new sorting algorithm is given and compared with merge sort implementation based on tape merging and binary merging algorithms. In addition, the parallel version of the proposed algorithm is implemented with openMP and pthreads frameworks.