Neural network accelerator

Abstract

The invention relates to the technical field of artificial intelligence, and provides a neural network accelerator. According to the neural network accelerator, simplified design is carried out by taking a layer as a basic unit; all the modules can run in parallel to achieve simultaneous running of all layers of the neural network on hardware so as to greatly improve the processing speed, a main control module can complete cutting distribution of operation tasks and corresponding data of all the layers, and then all the modules execute the operation tasks of the corresponding layers. Moreover,a three-dimensional multiply-accumulate array is also realized in the neural network accelerator architecture; therefore, multiplication and accumulation operations can be operated in parallel on a large scale; the convolution calculation efficiency is effectively improved; data multiplexing and efficient circulation of convolution operation parameters and convolution operation data among the modules are realized through a first cache region and a second cache region, data multiplexing between layers of the neural network is realized to reduce external data access, so that the power consumption is reduced while the data processing efficiency is improved, and the neural network accelerator can improve the computing power consumption ratio.

Description

technical field [0001] This application relates to the technical field of artificial intelligence, in particular to a neural network accelerator. Background technique [0002] With the explosion of convolutional neural network applications, the demand for neural network accelerators with high performance and low power consumption is becoming stronger and stronger. However, in the actual measurement process of the current neural network accelerators, it is found that the measured computing power of these neural network accelerators is usually much lower than the peak computing power of the target, and the power consumption during operation is too large, and the computing power and power consumption are relatively low. The technical problems affect its application on the embedded platform side, for example. Contents of the invention [0003] Based on this, it is necessary to provide a neural network accelerator to address the technical problem of relatively low computing po...

AI Technical Summary

Problems solved by technology

[0003] Based on this, it is necessary to provide a neural network accelerator for the technical problem of relatively low computing power and power consumption in traditional neural network accelerators.

Images