Hardware And Software Collaborative Design Of Hardware And Software Technology Research

Hardware/software co-design methodology is a new idea and method to solve theproblem in the development of embedded system, while hardware/softwarepartitioning is one key technology of hardware/software co-design. In this paper,knowledge of hardware/software partitioning has been introduced, several commoncalculation model of hardware/software partitioning was analyzed and then a directedacyclic graph (DAG) model was constructed in this paper. For the current mainstreamsingle processor target system architecture, in order to solve hardware/softwarepartitioning problem efficiently, in the case of considering the system softwarefunction module and hardware function module and hardware function modules canbe executed in parallel, critical path scheduling algorithm is proposed, the mutationstrategy of genetic algorithm is introduced into the discrete particle swarmoptimization algorithm, adaptive and efficient particle swarm algorithm was proposedfor hardware/software partitioning, adaptive mutation particle swarm algorithm hasbetter convergence and improving the stability and solution quality ofhardware/software partitioning algorithm.In the case of single target cost design constraint of the system partitioning,experimental results show that the algorithm proposed in the paper can produce bettersolution than first-come first-served (FCFS) scheduling algorithm and theimprovement is up to11%.For multi-objective partitioning problem, the proposedalgorithm was improved and the improved algorithm is applied to a given systemdesign constraints, the results is better than first-come first-served (FCFS) schedulingalgorithm, getting some research.