ILP based data parallel multi-task mapping/scheduling technique for MPSoC

Abstract

In this paper, we propose a multi-task mapping/scheduling technique for heterogeneous and scalable MPSoC. To cope with ever increasing computational demand and to utilize large number of cores effectively, the proposed technique considers data parallelism as well as task parallelism. The solution is based on ILP (Integer Linear Programming) and its effectiveness is proven by experiments with real-life examples.