Research On Energy Consumption Model And Performance Of Embedded Software Based On Architecture Level

With the rapid development and extensive applications of embedded systems, the low-power design of embedded systems becomes the hot topic. How to design a set of methods that have low energy consumption and high performance becomes the research hotspot nowadays. This paper analyses the fundamental factors which influence the energy consumption of software eigenvalue in the embedded systems firstly, and then set up a energy-consumption Petri net. According to the energy-consumption Petri net some component modules are built, such as the communication module and the hardware component module at the software architecture level of embedded system. Finally, the software attributes are analyzed and modeled by the energy consumption.In order to measure power consumption more accurately, this paper have analyzed how the algorithm and architecture design affects the system energy consumption in embedded system software, and the differences of algorithms in energy consumption comes from running time, besides the running time of algorithm decides the CPU usage. So the CPU usage depends on the algorithm. The difference of energy consumption of architecture design mainly derives from the component design, the number of components and the coupling degree of components, that is to say these three aspects decide the size of the energy consumption. Based on the above analysis, this paper introduces the execution time and the changes of the energy consumption function during this time, so that the energy-consumption Petri net is established. By using the energy-consumption Petri net model, the energy consumption model of component modules and the communication interface of each module are established. The energy consumption model of communication interfaces is used to connect the modules and form the model of energy consumption of the whole system.In order to avert influencing the reliability of software, after the energy consumption of the system model is established, this paper evaluates the software performance based on the embedded software architecture model of energy consumption. The corresponding measuring methods of the software reliability, real-time performance and security are given, and then the software interaction behavior is modeled according to the above measure methods.For all the above studies, this paper selects the embedded intelligent household system to implement the experimental verification on. The experiment results show that the embedded system software energy consumption model proposed in this paper has higher accuracy, and provides strong help for the analysis of software performance when designing the system. Therefore, the method proposed in this paper can effectively describe and analyze the energy consumption problems of embedded software. What is more, this method can provide strong help for the design of low-power and high-performance software in embedded systems.