Time domain in-memory computing array structure based on magnetic random access memory

Abstract

The invention discloses a time domain in-memory computing array structure based on a magnetic random access memory (MRAM), and belongs to the field of integrated circuit design. The circuit is characterized in that the structure comprises a dual-mode storage array, a self-adaptive pipeline decoder, a pre-charging circuit, a column selector, a sensitive amplifier, an input and output unit, a delaydifference quantization unit, a counting unit, a time sequence control circuit and a mode selection module. The method has a standard read-write mode and an in-memory calculation mode. Under the standard read-write mode, the read-write operation of the data in the storage array can be realized; the in-memory calculation mode can realize multiply-accumulate operation in binary neural network calculation. Multiply-accumulate calculation is completed during the reading of data; meanwhile, the delay quantization unit and the storage array are integrated together to reduce memory access energy consumption; compared with a conventional Von Noemann architecture neural network accelerator, the network operation energy efficiency is effectively improved.

Description

technical field [0001] The invention relates to the field of integrated circuit design, in particular to a magnetic random access memory (MRAM)-based time-domain in-memory computing array structure and a high-energy-efficiency circuit design method for realizing binary neural network convolution calculation based on the memory. Background technique [0002] In recent years, Convolutional Neural Networks (CNN) have shined in areas such as image recognition, leading a new wave of artificial intelligence. Convolutional neural network is a hierarchical network structure, such as figure 1 As shown, it mainly includes the following hierarchical structures: data input layer, convolution calculation layer, function excitation layer, pooling layer, and fully connected layer. Such as figure 2 As shown, the calculation process can be summarized as that the current network layer is to weight and sum the activation values ​​of the previous layer, then add a bias item, and finally obta...

AI Technical Summary

Problems solved by technology

And based on the delay difference quantization unit disclosed in the present invention, the problem of low quantization accuracy of traditional in-memory calculations is solved

Images