Algorithm for Parallelization of Cubic Spline Computational Processes in Distributed Computing Systems

Abstract

The article is devoted to the creation of an algorithm for the parallelization of cubic-spline computational processes in distributed computing systems. The article considers the creation of an algorithm or parallelization of cubic-spline computational processes using MPI (Message Passing Technology) technology, studying the problems of parallel streams in a distributed computing system and parallel transmission and reception of messages in real-time on networked computers. Spline coefficients were calculated using point formulas that fully satisfy the parallelization requirements. Point formulas allow the implementation of time-consuming computational processes in parallel, as they do not require the solution of a system of equations. The acceleration coefficient and efficiency of the proposed parallel algorithm are calculated and presented in the form of a comparative table relative to the computational processes on several computers. A large number of geographic signals were selected as the incoming signal.