Storage and calculation integrated chip adder

Abstract

The invention relates to a storage and calculation integrated chip adder, in the flourishing development of artificial intelligence nowadays, a storage and calculation integrated chip has unique conditions for processing a neural network algorithm, and a fast multiplier based on an analog signal can rapidly realize various artificial intelligence algorithms. However, only a simple function is achieved when rapid operation of addition is achieved, large-scale calculation cannot be achieved, and large-scale addition operation becomes the bottleneck of the speed of a storage and calculation integrated chip. According to the invention, a mode of carrying out preprocessing on the MCU and combining the MCU with the storage and calculation integrated chip is provided, so that the defects of the storage and calculation integrated chip in addition are successfully overcome. The system comprises an MCU module and a storage and calculation integrated chip module. The MCU module preprocesses data, the storage and calculation integrated chip performs summator operation, and the MCU module and the storage and calculation integrated chip are mutually independent and perform own functions. When the adder is used, a single integer and floating-point number adder or a vector shaping adder and a vector floating-point number adder can be selected according to the type of the adder.

Description

technical field [0001] The invention relates to the technical field of high-performance computing, in particular to a memory-computing integrated chip adder. Background technique [0002] The adder is the basic unit in the computing device. It has a specific implementation method in CPU, GPU, MCU, FPGA, etc. The main implementation method is to split the addition of different types of data into binary logic operations. Adders are often used in computer arithmetic and logic components, performing logic operations, shifting and instruction calls, etc. In electronics, an adder is a digital circuit that performs digital addition operations. [0003] The rapid development of memory-computing integrated chip technology in recent years has created opportunities for high-speed real-time operation of matrix operations. With the rise of application fields such as the Internet of Things and artificial intelligence, technology has been extensively researched and applied by academic ci...

AI Technical Summary

Problems solved by technology

[0005] In today's vigorous development of artificial intelligence, the memory-computing integrated chip has unique conditions for processing algorithms such as neural networks. Its fast multiplier based on analog signals can quickly implement various artificial intelligence algorithms. It only has simple functions and cannot achieve large-scale calculations. The large-scale addition operation has become the bottleneck of the speed of the memory-calculation integrated chip. The main problem is that it cannot increase the operation speed while ensuring the accuracy.

Images