Neural network computing device and processor comprising same

Abstract

The invention provides a neural network computing device and a processor comprising the same, wherein the computing device comprises a pulsating array processing unit and a main processor. The main processor is used for controlling the loading of computational elements in a neural network to the pulsating array processing unit and the transmitting in the pulsating array processing unit; the pulsating array processing unit is composed of a plurality of processing units; each processing unit performs calculations on received computational elements and / or passes the received computational elements to the next processing unit, wherein the computational elements include neuron data and corresponding weight values. The computing device of the invention is capable of speeding up the computational speed of the neural network and reducing the demand for bandwidth in the computing process.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a neural network computing device and a neural network processor including the computing device. Background technique [0002] Deep learning is an important branch of the field of machine learning, which has made major breakthroughs in recent years. The neural network model trained with deep learning algorithms has achieved remarkable results in image recognition, speech processing, intelligent robots and other application fields. [0003] The deep neural network simulates the neural connection structure of the human brain by building a model, and describes the data features hierarchically through multiple transformation stages when processing signals such as images, sounds, and texts. With the continuous improvement of the complexity of the neural network, the neural network technology has many problems in the actual application process, such as occupying a lot ...

AI Technical Summary

Problems solved by technology

[0004] However, since the neural network processor is a computationally intensive and memory-intensive processor, on the one hand, the neural network model includes a large number of multiplication and addition operations and other nonlinear operations, which require the neural network processor to maintain high load operation to ensure the neural network The computing requirements of the network model; on the other hand, there are a large number of parameter iterations in the neural network computing process, and the computing unit needs to access a lot of memory, which greatly increases the bandwidth design requirements of the processor and increases memory access power consumption.

[0005] Therefore, it is necessary to improve the existing neural network processors to solve the problems of high computing circuit hardware overhead and large on-chip memory access bandwidth.

Images