Mapping Applications to Tiled Multiprocessor Embedded Systems

Abstract

Modern multiprocessor embedded systems execute a large number of tasks on shared processors and handle their complex communications on shared communication networks. Traditional methods from the HW /SW codesign or general purpose computing domain cannot be applied any more to cope with this new class of complex systems. To overcome this problem, a framework called Distributed Operation Layer (DOL) is proposed that enables the efficient execution of parallel applications on multiprocessor platforms. Two main services are offered by the DOL: systemlevel performance analysis and multi-objective algorithmarchitecture mapping. This paper presents the basic principles of the DOL, the specification mechanisms for applications, platform and mapping as well as its internal analytic performance evaluation framework. To illustrate the presented concepts, an MPEG -2 decoder case study is presented.