Brain-like chip and electronic equipment

Abstract

The invention discloses a brain-like chip and electronic equipment. A traditional artificial intelligence chip faces the dilemma of a Moore wall, a memory wall and an energy efficiency wall, compared with the prior art, although the calculation energy efficiency of most brain-like chips is obviously improved, the traditional artificial intelligence chip still faces the problems of an integrated wall and an energy efficiency wall at present, and the requirements for cost and power consumption in an edge calculation scene cannot be substantially met. In order to solve the problems of single cost, power consumption and task capability, multiple types of special spiking neural network cores are integrated in a chip, and the network can call special cores with different resource scales and characteristics, so that the power consumption and the cost are remarkably improved, and the network performance is improved. And the performance of the network can be improved through multi-network parallel processing of multi-modal information and collaborative decision. The multi-task parallel processing capability enables the brain-like chip disclosed by the invention to be more similar to a biological brain, and the brain-like chip disclosed by the invention is more intelligent, low in cost and high in energy efficiency, and provides possibility for intelligent interconnection of all things. The method is suitable for the fields of brain-like chips and AIoT.

Description

technical field [0001] The invention relates to a brain-inspired chip and electronic equipment, and in particular to a multi-mode brain-inspired chip and electronic equipment with multiple dedicated neural network cores. Background technique [0002] The three most important indicators of artificial intelligence (AI) chips are cost, power consumption and performance. At present, the cost and power consumption of traditional AI chips based on artificial neural network (ANN) are relatively high, and it is difficult to meet the needs of edge computing, so people began to turn their attention to neuromorphic computing inspired by the brain. [0003] Neuromorphic (also known as neuromorphic) computing is a new type of computing architecture developed in recent years. By running spiking neural networks (SNNs) on neuromorphic hardware (commonly known as brain-like chips), real-time reasoning of input signals can be achieved. [0004] IBM's True-North is a typical representative o...

AI Technical Summary

Problems solved by technology

[0005] But the redundancy of this scheme is still too high, refer to image 3 , the neuron integration level of the chip is usually difficult to break through 3k / mm 2

In order to improve neuron integration / reduce silicon cost, one technical approach is to alleviate through advanced chip process nodes, such as Loihi-2 chips, but this does not essentially solve the problem

In addition, the known brain-like chips can only deploy a single spiking neural network, and can only meet the needs of a single type of task, and the chip function is relatively single

Images