Embedded software energy consumption testing method based on improved neural network

Abstract

The invention relates to an embedded software energy consumption testing method based on an improved neural network, and belongs to the field of embedding technology. The method comprises the following steps: analyzing introduced embedded software to obtain a node relation diagram; analyzing the clustering relation, out-degree and in-degree of each node according to the node relation diagram, and calculating a clustering relation average value according to the clustering relation of each node; substituting the clustering relations, out-degree values and in-degree values of input nodes and the clustering relation average value of the clustering relation of each input node into the improved neural network; calculating the error of the output value of each node according to the output value of each output node and the set training target value of each output node; performing energy consumption testing on embedded software corresponding to the number of optimized nodes in a Linux system by adopting a power consumption emulator to obtain an optimized energy consumption value. By adopting the method, the code reusability is improved, and energy consumption optimization is realized.

Description

technical field [0001] The invention relates to an embedded software energy consumption testing method based on an improved neural network, and belongs to the field of embedded technology. Background technique [0002] Embedded software energy consumption is one of the hot issues in the current embedded field. In order to optimize software energy consumption, many testing and optimization methods have emerged. Energy consumption analysis is usually based on instruction-level analysis or at the algorithm and software architecture level. [0003] However, in the energy consumption test through neural network training, people only use the traditional neural network model to train the feature quantities in the software, but these feature quantities cannot accurately analyze the energy consumption of the software. consumption analysis and research. An embedded software energy consumption testing method for improved neural network, which analyzes the node relationship diagram in...

AI Technical Summary

Problems solved by technology

[0003] However, in the energy consumption test through neural network training, people only use the traditional neural network model to train the feature quantities in the software, but these feature quantities cannot accurately analyze the energy consumption of the software. consumption analysis and research

Images