This paper presents an efficient run-time Hardware/Software scheduling approach. This scheduling heuristic consists in mapping on-line the different tasks of a highly dynamic application in such a way that the total execution time is minimized. Our approach takes advantage of the reconfiguration property of the considered architecture to adapt processing to the system dynamics. We compare our heuristic with another similar approach. We present the results of our scheduling method on an image processing application. Our experiments include simulation and synthesis results on a Virtex2P based platform. This results show a better performance against existing methods.
展开▼
