Encoding type flash memory device and encoding method

Abstract

The invention discloses an encoding type flash memory device and an encoding method. The encoding type flash memory device comprises at least one flash memory array structure unit, a plurality of comparators and a plurality of adders, wherein each flash memory array structure unit in at least one flash memory array structure unit is a 3D NAND FLASH array structure unit and is used for realizing encoding operation to generate a source line voltage on each source line in a plurality of source lines of the flash memory array structure unit; each comparator in the plurality of comparators is correspondingly connected with each source line and is used for converting the voltage of the source line of each correspondingly connected source line into an output result in a binary form; and each adder in the plurality of adders is connected with at least two comparators in the plurality of comparators through each corresponding source line, and is used for carrying out summation operation on at least two output results corresponding to at least two comparators. According to the encoding type flash memory device and the encoding method, efficient and accurate full connection layer or convolutional layer operation can be realized, so a deep neural network is realized.

Description

technical field [0001] The invention relates to the technical field of semiconductor devices and integrated circuits, in particular to an encoding flash memory device and an encoding method for realizing a deep neural network. Background technique [0002] Deep neural networks are currently widely used in image processing and speech recognition, and have shown excellent performance. The deep neural networks in the prior art mostly rely on the traditional von Neumann architecture for calculation, but limited by the separation of storage units and computing units in the von Neumann computing architecture, data transmission through the bus will cause delays Increase and energy consumption increase, so that there is a bottleneck in the improvement of the data processing capability and energy efficiency ratio of the deep neural network. For this reason, the prior art further proposes an encoded flash memory device that performs analog operations based on the 3D NAND FLASH struct...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the technical problems of low calculation accuracy, low energy efficiency ratio, non-maximized computing architecture efficiency, and unsaturated parallelism in the encoded flash memory device that performs analog operations based on the 3D NAND FLASH structure in the prior art, the present invention provides An encoded flash memory device and encoding method

Images