Convolution neural network (CNN) hardware accelerator and acceleration method

Abstract

The invention discloses a convolution neural network (CNN) hardware accelerator and an acceleration method. The accelerator comprises an input buffer and a plurality of operation units , and is characterized in that the input buffer is used for caching input feature picture data, the plurality of operation units respectively share the input feature picture data to perform a CNN convolution operation, each operation unit comprises a convolution kernel buffer, an output buffer and a multiplier-adder unit formed by a plurality of MAC components, the convolution kernel buffer receives convolutionkernel data returned from an external storage component, the convolution kernel data is provided for each MAC component of the multiplier-adder unit, each MAC component receives the input feature picture data and the convolution kernel data to perform a multiply accumulation operation, and an intermediate result of the operation is written into the output buffer. The acceleration method is a method applying the accelerator. The CNN hardware accelerator and the acceleration method can improve the CNN hardware acceleration performance, and have the advantages of high data reuse rate and efficiency, small amount of data migration, good expansibility, small bandwidth required by the system, small hardware overhead and the like.

Description

technical field [0001] The present invention relates to the technical field of deep learning (Deep Learning, DL) hardware acceleration, in particular to a CNN (Convolution Neural Networks, Convolution Neural Networks) hardware accelerator and an acceleration method. Background technique [0002] Convolutional neural network (CNN) is one of the most important algorithms in deep learning. It is widely used in many fields such as target recognition, unmanned driving and artificial intelligence because of its high precision and small weight. In each network layer of convolutional neural network CNN, the convolutional layer accounts for more than 90% of the calculation amount and computing time of the entire network. Accelerating the operation of the convolutional layer is the key to improving the performance of CNN. Therefore, it is urgent to design a CNN hardware accelerator to Improve the efficiency of convolution operations. [0003] The CNN convolutional layer algorithm has...

AI Technical Summary

Problems solved by technology

[0005] 1) When the pulsating array structure is adopted, the computing unit (PE) is prone to a high idle rate, which in turn reduces the efficiency of the accelerator;

[0006] 2) When the systolic array structure is used, the scalability of the systolic array is poor due to the proportional increase in bandwidth required to maintain the required acceleration factor

Images