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

Multi-GPU-based graph data processing system and method

A technology for processing systems and graph data, applied in the field of graph data processing systems based on multi-GPU, can solve the problems of high communication overhead, limited size of graph data that can be processed, low task parallelism, etc., to improve system performance and reduce synchronization waits overhead, the effect of an efficient programming framework

Active Publication Date: 2017-09-01
HUAZHONG UNIV OF SCI & TECH
View PDF6 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This solves the technical problems of high communication overhead between GPU-CPU, low task parallelism and limited scale that can process graph data in the GPU-based graph system in the prior art.

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
  • Multi-GPU-based graph data processing system and method
  • Multi-GPU-based graph data processing system and method
  • Multi-GPU-based graph data processing system and method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not constitute a conflict with each other.

[0043] The present invention adopts the idea of ​​using computing time to cover up communication time, and solves the problems of high communication overhead between CPU and GPU and limited scale of processing graph data in the current GPU-based graph system, and makes full use of the parallel computing capability of GPU and CPU computing The ability to accelerate the process of graph data proce...

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 discloses a multi-GPU-based graph data processing system and method. The method comprises the steps of firstly dividing input original graph data through a data division module, and allocating the divided data to GPU devices; secondly allocating different computing tasks to vertexes on the GPU devices according to computing attributes through a task management module, and managing an executive sequence of various tasks; thirdly finishing execution of parallel computing tasks on the GPU devices while ensuring data consistency through an executive module; and finally finishing a data collection process after computing completion through a data collection module. According to the system and the method, a new task allocation and management mechanism is proposed; parallel efficiency is maximized; and communication is masked with computing, so that the problems of high GPU-CPU communication overhead and finite graph data processing scale of an existing GPU-based graph processing system are solved and the efficiency of processing large-scale graph data in a GPU is improved.

Description

technical field [0001] The invention belongs to the technical field of graph data processing and high-performance computing, and more particularly relates to a multi-GPU-based graph data processing system and method. Background technique [0002] At present, many practical applications involve the processing of graph data. The traditional stand-alone graph system mainly improves the efficiency of large-scale graph data processing by means of data partitioning, optimizing memory access efficiency and reducing disk I / O overhead, but there are still parallel problems. Low computing efficiency, insufficient disk bandwidth utilization, and low memory access efficiency. [0003] The GraphChi system divides large-scale graph data into small data blocks, and proposes the concept of Shards and the mechanism of parallel sliding windows. While improving the continuity of memory data access, shard blocks are transferred on demand, reducing random disk access. However, due to the fragme...

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 Applications(China)
IPC IPC(8): G06F9/50G06F9/52G06F9/54
CPCG06F9/5038G06F9/52G06F9/546
Inventor 石宣化金海罗璇赵鹏
Owner HUAZHONG UNIV OF SCI & 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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com