Data communication method and system and spiking neural network operation system

Abstract

The invention discloses a data communication method and system and a spiking neural network operation system, and the data communication system comprises a first interface module, a data frame buffer module, a second interface module and a control module. The first interface module is used for receiving data transmitted by an upper computer through the first bus, converting downlink data frames, reading uplink data frames, converting the uplink data frames and sending the uplink data frames to the upper computer through the first bus. The data frame caching module is used for caching downlink data frames and uplink data frames; the second interface module is used for reading the downlink data frame, converting the downlink data frame into a pulse according to a chip protocol, transmitting the pulse to the pulse neural network chip, and converting the received pulse into an uplink data frame according to the chip protocol; the control module is used for controlling data transmission interaction between the first interface module and the upper computer and data transmission interaction between the second interface module and the pulse neural network chip. According to the invention, data transmission of large data volume can be realized.

Description

technical field [0001] The present invention relates to the technical field of integrated circuits, in particular to a data communication method and system, and an impulse neural network computing system. Background technique [0002] Spiking Neural Network (SNN) draws on the information processing mechanism of the human brain and simulates the working mode of the human brain. It directly uses the pulse firing time of neurons as its input and output, makes full use of time information, and has pulse sparsity. At the same time, when the spiking neuron receives the pulse, it will accumulate the input to the membrane potential, and the pulse will be emitted when the membrane potential reaches the threshold, which is event-driven. Since SNN has pulse sparsity and itself is event-driven, it can process information more efficiently similar to the human brain, achieving low latency and low power consumption. [0003] The existing spiking neural network computing system is mainly c...

AI Technical Summary

Problems solved by technology

However, when the input size of the spiking neural network is large, tens of thousands of data frames need to be transmitted in each time step of the inference operation (at a working frequency of 192Mhz, a time step is about 6.25 milliseconds), the above system will have data overflow. So there is a problem of data loss

Images