Method for accelerating convolution neutral network hardware and AXI bus IP core thereof

Abstract

The invention discloses a method for accelerating convolution neutral network hardware and an AXI bus IP core thereof. The method comprises the first step of performing operation and converting a convolution layer into matrix multiplication of a matrix A with m lines and K columns and a matrix B with K lines and n columns; the second step of dividing the matrix result into matrix subblocks with m lines and n columns; the third step of starting a matrix multiplier to prefetch the operation number of the matrix subblocks; and the fourth step of causing the matrix multiplier to execute the calculation of the matrix subblocks and writing the result back to a main memory. The IP core comprises an AXI bus interface module, a prefetching unit, a flow mapper and a matrix multiplier. The matrix multiplier comprises a chain type DMA and a processing unit array, the processing unit array is composed of a plurality of processing units through chain structure arrangement, and the processing unit of a chain head is connected with the chain type DMA. The method can support various convolution neutral network structures and has the advantages of high calculation efficiency and performance, less requirements for on-chip storage resources and off-chip storage bandwidth, small in communication overhead, convenience in unit component upgrading and improvement and good universality.

Description

technical field [0001] The invention relates to a hardware acceleration technology of a convolutional neural network, in particular to a hardware acceleration method of a convolutional neural network and an AXI bus IP core thereof. Background technique [0002] The core challenge of the next generation of smart device processors is to be able to perceive and understand the human world to provide an ecological environment that enhances user experience and connects user preferences, and can interact with users similar to humans. Convolution Neural Network (CNN) is one of the perception models representing the most advanced development level at present. This model can parse the original input data into symbols layer by layer and extract complex multi-layer combination features. Machine vision and auditory systems have achieved great success and wide application. In 2013, MIT Technology Review magazine ranked deep learning represented by convolutional neural network as the top t...

AI Technical Summary

Problems solved by technology

ASIC casting requires a lot of capital and human resources, and cannot be modified

Images