The Analysis of Energy Performance in Use Parallel Merge Sort Algorithms

Abstract

The issue of productivity and energy is an important objective of the optimization of parallel applications. The
size of the problem for a large number of data on multiprocessor platforms forces the use of parallel algorithms.
Efficient management of large memories using modern processors in Big data processing requires innovative
techniques and efficient algorithms. For years have found the results of tests conducted on methods for use
in various computing environments and improvements. This article shows the energy consumption analysis
by parallel sorting algorithms. Sort algorithms are used in information systems and databases, to select and
organize the information. The subject of this article is research into energy consumption and computational
complexity for parallel sorting methods by merging compared to classic methods. The tests carried out confirm
the reduction of energy consumption by using parallel sorting algorithms. The presented parallel fast sort and
parallel modified merge sort for large task dimensions have less power consumption than classic methods and
can be used successfully in NoSQL databases.