Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397

Issue N5 2014 year

Improving the Efficiency of Parallel Programs by the Time They Run through Judicious Placement of Data in a Distributed Memory Calculator
V. M. Bakanov , e-mail: vbakanov@hse.ru, V. V. Palagin

This paper discusses the issues of effective distribution of data across local ethyl resources (RAM) Multiprocessor Computing Systems (MVS) MPP architecture to minimize downtime calculators during communication between nodes MVS. The formal model of the process of implementing such a distribution, source-level programs and the technique, optimization.

Keywords: parallel programming, architecture MPP, the distribution of data across local storage resources of compute nodes, selects the optimal (rational) distribution
pp. 6–13