Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for realizing neural network calculation by using cell array computing system

A cell array and neural network technology, applied in the field of computers and computer applications, can solve problems such as affecting the overall performance of the computer and detrimental to the efficient implementation of neural network calculations, and achieve the effect of speeding up the operation speed, improving the overall performance, and having a good cost-effectiveness ratio.

Active Publication Date: 2016-07-06
SHANGHAI CIYU INFORMATION TECH
View PDF5 Cites 23 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0010] The problem to be solved by the present invention is that the computer architecture in the prior art affects the overall performance of the computer due to the communication bottleneck between the CPU, memory and storage, which is not conducive to the efficient realization of neural network calculations

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Method for realizing neural network calculation by using cell array computing system
  • Method for realizing neural network calculation by using cell array computing system
  • Method for realizing neural network calculation by using cell array computing system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] Due to the communication bottleneck between the CPU, the memory and the storage in the computer architecture of the prior art, the improvement of the overall performance of the computer is affected, and the cost-effectiveness is relatively poor.

[0044] After research, the inventor of the present application considers that if the three functions of memory, storage and calculation are integrated into one chip, a relatively simple unit with independent calculation and storage functions will be formed, and a large number of such units will form a dense network. Communication network, in order to realize the function of data broadcasting / group sending and the internal network capable of large-scale parallel transmission of data, a computing architecture similar to the human brain can be developed, which is equivalent to building a large number of microcomputers in one on chip.

[0045] To this end, the technical solution of the present invention provides a computing archit...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A method for implementing neural network calculations using a cell array computing system, the cell array computing system comprising: a main control CPU, a cell array, and a cell array bus; dimensional array; the main control CPU communicates with each cell in the cell array through the cell array bus; there is a communication interface between adjacent cells, and data can be sent to each other; the method includes: store one or more neuron function codes; any cell participating in the neural network calculation receives the execution result of the neuron function code output from other cells, and executes the code stored in the cell as the input data of the cell The code of the neuron function, and output the execution result, or store it in the preset address. The invention can overcome the communication bottleneck between the CPU, memory and storage in the existing computer architecture, and improve the overall performance of the system.

Description

technical field [0001] The invention relates to the technical field of computers and computer applications, in particular to a method for realizing neural network calculations using a cell array computing system. Background technique [0002] Generally speaking, a computer mainly includes three core parts: a central processing unit (CPU, Central Processing Unit), memory and storage. [0003] Through the unremitting efforts of some of the world's top companies, the CPU has evolved into an extremely complex semiconductor chip. The number of MOS tubes inside the top CPU core can exceed 100 million. The current industry trend is that it is difficult to increase the operating frequency of the CPU due to power consumption constraints. It is also difficult to improve the operating efficiency of the already extremely complex modern CPU. New CPU products are increasingly evolving towards multi-core. [0004] In terms of memory, the current dominant technology is Dynamic Random Ac...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06N3/00G06F15/163
Inventor 戴瑾郭一民王践识
Owner SHANGHAI CIYU INFORMATION TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products