Method and device for data processing, DMA controller, and computer readable storage medium

Abstract

Provided are a method and a device for data processing, a DMA controller, and a computer readable storage medium. The method comprises: acquiring feature information and parameter information of an original input feature map; generating second DMA configuration information according to the feature information, and generating first DMA configuration information and third DMA configuration information according to the feature information and the parameter information; constructing a target input feature map according to the first DMA configuration information; and reading input data from from the original input feature map according to the second DMA configuration information; storing the input data to the target input feature map according to the third DMA configuration information. By applying the embodiment of the present invention, data transit in CNN can be implemented by the DMA controller, and a CPU does not need to implement data transit in the CNN, thereby reducing CPU load, transitting the data more efficiently, and thereby accelerating CNN operation without losing flexibility.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a data processing method, equipment, DMA (Direct Memory Access, direct memory access) controller and a computer-readable storage medium. Background technique [0002] In machine learning, CNN (Convolutional Neural Network, Convolutional Neural Network) is a feedforward neural network whose artificial neurons can respond to surrounding units within a part of the coverage area, and it has excellent performance for large-scale image processing. CNN is a multi-layer neural network, each layer is composed of multiple two-dimensional planes, and each plane is composed of multiple independent neurons. Generally, CNN can be composed of a convolutional layer and a pooling layer. The function of the convolutional layer is to extract various features of the image, and the function of the pooling layer is to perform two feature extractions on the original feature signal to r...

AI Technical Summary

Problems solved by technology

For example, image algorithms involve operations of fixed matrices, such as Gaussian (Gaussian) matrices for Gaussian filtering. When the CPU completes matrix operations, it also needs to move data, which increases the burden on the CPU.

Images