TASKS MAPPING METHOD FOR COARSE GRAIN RECONFIGURABLE SYSTEMS

This paper deals with analysis of existing approaches to tasks mapping on reconfigurable computing systems with special attention paid to mapping methods for coarse grained reconfigurable computing systems. The purpose and objectives of a new heuristic method for tasks mapping on coarse grained re...

Full description

Bibliographic Details
Main Author: A. S. Rumyantsev
Format: Article
Language:English
Published: Saint Petersburg National Research University of Information Technologies, Mechanics and Optics (ITMO University) 2014-01-01
Series:Naučno-tehničeskij Vestnik Informacionnyh Tehnologij, Mehaniki i Optiki
Subjects:
Online Access:http://ntv.ifmo.ru/file/article/8398.pdf
Description
Summary:This paper deals with analysis of existing approaches to tasks mapping on reconfigurable computing systems with special attention paid to mapping methods for coarse grained reconfigurable computing systems. The purpose and objectives of a new heuristic method for tasks mapping on coarse grained reconfigurable computing systems are produced on the base of the carried out analysis. This novel method for tasks mapping on coarse grained reconfigurable computing systems is based on the method of graph partitioning with pushing vertices, graph covering algorithm, a heuristic approach to optimizing and packaging for a particular graph on coarse grained reconfigurable computing system and displaying methods of data flow graph on resources of coarse grained reconfigurable computing system. The simulation was conducted on system model with coarse grained reconfigurable hardware accelerator MATRIX. Experimental results are given. They prove the effectiveness of the proposed approach as compared with the widely used methods for tasks mapping on coarse grained reconfigurable computing systems and the ability to use dynamic functional parameters of coarse grained reconfigurable computing system to further improvement of the mapping results.
ISSN:2226-1494
2500-0373