Integrated pipelined task scheduling and core mapping for streaming applications on multi-core systems

In this paper, we present an efficient integrated pipelined task scheduling and core mapping approach called IPTSCM for streaming applications on NoC-based heterogeneous multi-core architecture. First, a polyhedron model is developed to formulate the integrated pipelined scheduling and mapping problem. Second, a series of constraints are applied, which reduce the solution space greatly. Third, an optimized simulated annealing algorithm is adopted to find the optimal solution in the space. Experiments demonstrate our method can yield more than 3.02 times higher throughput and 11.63% lower energy consumption on average compared with the existing algorithm.