SCHEME OF OPERATIONS ENLARGING FOR PARALLELIZED ALGORITHMS

One of the solutions of the problem designing the effective numerical methods for parallel architecture computers is building the information graph by serial algorithm and transforming by list consecutions method to parallel form with sufficient edge width and implementation with parallel system. This article points out that it is possible not only to formalize this process but also to find out the optimal solution for parallelizing the algorithm with a glance to such parameters as a number of processor units, computation time and computation density per unit time

Authors: M. S. Kupriyanov, Yu. A. Shichkina

Direction: Informatics, management and Computer Technology

Keywords: Parallel algorithm, information graph, width of graph, list consecutions


View full article