Research on Low Power Optimization Technology of Power ASIC Based on a Neural Network Algorithm

To improve the low-power optimization effect of power ASIC, a low-power optimization technology of power ASIC based on a neural network algorithm is designed in this paper. This paper analyzes the neural network algorithm, gives the calculation process of the convolution layer, uses the neural network algorithm to calculate the cycle flow, constructs the migration model, designs the low-power optimization equilibrium model of the integrated circuit according to CNN, and realizes the low-power optimization of power ASIC. The experimental results show that the average optimization effect of this method is 8.9%, and the prediction time of integrated circuit low power consumption is only 1.8 s, which shows that this method can effectively improve the prediction efficiency and has good power optimization.