Research On Complex Networks Based Embedded Software Power Optimization

Nowadays, with the development of mobile device, and popularity of the Embedded Equipment, microprocessor craftsmanship promotion and microprocessor frequency rising, embedded system power optimization has been becoming a very urgent problem with limit power supply in an embedded system, which embedded system designers are facing and eager to solve. Hence, a new power optimization model based on embedded software is proposed to solve power dissipation problems.Embedded software is playing a more and more important role in an embedded system. Power optimization based on embedded software is becoming more important than before. Power optimization is proposed based on DFG graph(Digital flow graph), generated from program. With the help of complex networks, two algorithms are put forward to implement power optimization.Firstly, the shortages in current researches are pointed out based on reviewing many literatures on embedded software power optimization. Many power optimization models have been issued to solve embedded software power optimization, such as dynamic voltage and frequency, static power, cold scheduling, loop optimization and inline optimization.Secondly, two algorithms based on complex networks are proposed to solve embedded system power optimization. The proof which DFG graph fits to complex networks is given in the third section. Then the characteristics of complex networks are applied to the DFG graph, to calculate the in-out degree, according to the in-out degree, a waiting queue is formed, then switch algorithm is applied to the queue to exchange the register nodes and memory nodes.The sparc project is carrying through in the lab which is the author in. Finally, the two algorithms are finally integrated into the sparc project, so as to the spare project can run at a lower power consumption.