Hardware software partitioning with integrated hardware designspace exploration

摘要

This paper presents an integrated approach to hardware softwarepartitioning and hardware design space exploration. We propose a geneticalgorithm which performs hardware software partitioning on a task graphwhile simultaneously contemplating various design alternatives for tasksmapped to hardware. We primarily deal with data dominated designstypically found in digital signal processing and image processingapplications. A detailed description of various genetic operators ispresented. We provide results to illustrate the effectiveness of ourintegrated methodology