Hardware/Software Partitioning Based on Hybrid Genetic and Tabu Search in the Dynamically Reconfigurable System

This paper proposes a hardware/software partitioning algorithm which can be applied to the dynamically reconfigurable system. Firstly, based on reconfigurable system structure, this paper brings forward a kind of system model and its task description ,then use genetic/tabu search (GATS) integration strategy, in the condition of resource constrain of the reconfigurable systems, specific applications will be mapped to software and hardware platform. Secondly, using configuration prefetching and scheduling strategy it will be found out that the shortest time assignment and the execution order of the entire task flow diagram in the partition result. The experiment results have shown that this method can effectively map the task graph to reconfigurable system, and is a kind of method with high performance.