Computing cluster and computing cluster configuration method

A computing cluster and computing module technology, applied in the computer field, can solve problems such as insufficient memory and computing resources

Active Publication Date: 2019-06-18
上海燧原智能科技有限公司 +1
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the memory and computing resources that a single computer can provide are far from enough

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
  • Computing cluster and computing cluster configuration method
  • Computing cluster and computing cluster configuration method
  • Computing cluster and computing cluster configuration method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The embodiment of the present invention discloses a computing cluster to provide a solution for training a deep neural network.

[0038] The topological structure of the above computing cluster corresponds to the topological structure of the deep neural network to be trained, so as to form different topological structures according to the application requirements of deep learning.

[0039] figure 1 The topological correspondence between computing clusters and deep neural networks is shown.

[0040] The number of single-layer networks A contained in the deep neural network to be trained is N (N is a natural number). Wherein, each single-layer network A includes multiple functional nodes a.

[0041] The computing cluster includes: a sub-computing cluster C corresponding one-to-one to the above-mentioned N-layer single-layer network.

[0042] For the convenience of naming, any single-layer network A can be expressed as the i-th layer network (A i ), i is less than or e...

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

The invention provides a computing cluster for training a deep neural network and a computing cluster configuration method. Wherein, the number of single-layer networks included in the deep neural network to be trained is N, and N is a natural number; wherein, each single-layer network contains multiple functional nodes; any single-layer network is represented as the i-th layer network; i is less than or equal to N; the computing cluster includes: a sub-computing cluster corresponding to the N-layer single-layer network; wherein, the sub-computing cluster corresponding to the i-th layer single-layer network is the i-th sub-computing cluster; the computing in the i-th sub-computing cluster Nodes correspond one-to-one to functional nodes in the i-th layer single-layer network; computing nodes include at least one basic computing module; each basic computing module includes m accelerator cards that are point-to-point fully interconnected, and each accelerator card includes n point-to-point fully interconnected artificial intelligence chips; m is a natural number not less than 2; n is a natural number; any basic computing module is interconnected with adjacent basic computing modules through chip-level point-to-point connections.

Description

technical field [0001] The invention relates to the field of computers, in particular to a computing cluster and a configuration method for the computing cluster. Background technique [0002] Deep neural networks are a popular topic in the field of machine learning in recent years and have achieved advanced results on many tasks. [0003] Deep neural networks need to be trained before they can be put to use. Training a neural network model requires a lot of memory and computing resources. Therefore, the memory and computing resources that a single computer can provide are far from enough. Contents of the invention [0004] In view of this, the present invention provides a computing cluster and a computing cluster configuration method for training a deep neural network. [0005] In order to achieve the above-mentioned purpose of the invention, the present invention provides the following technical solutions: [0006] A computing cluster for training deep neural network...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06F15/177G06N3/04G06N3/08H04L29/08
Inventor 江斌刘彦李翔胡维
Owner 上海燧原智能科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products