Co-processing acceleration method, apparatus, and system

Active Publication Date: 2013-06-13
HUAWEI TECH CO LTD
View PDF0 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Benefits of technology

[0020]In the embodiments of the present invention, through the foregoing technical solutions, a public buffer card is used as a public data buffer channel between each compute node and each co-processor card of a computer system, and to-be-processed data does not need to be

Problems solved by technology

However, the processing capability of a single computer is limited, and is likely to become a bottleneck of improving system performance in the foregoing large-scale computing scenarios, and this problem is effectively solved as a cluster system emerges.
By adopting

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
  • Co-processing acceleration method, apparatus, and system
  • Co-processing acceleration method, apparatus, and system
  • Co-processing acceleration method, apparatus, and system

Examples

Experimental program
Comparison scheme
Effect test

Example

[0068]In Embodiment 1 of the present invention, through the foregoing technical solution, according to a co-processing request message sent by each compute node in the computer system, to-be-processed data on which processing is requested by each compute node is allocated to the idle co-processor card in the system for processing. The compute node does not need to consume its own resources to perform allocation of the to-be-processed data, which reduces resource overheads of each compute node itself. The public buffer card is used as a public data buffer channel between each compute node and each co-processor card of the computer system, and the to-be-processed data does not need to be transferred by the memory of the compute node, which avoids overheads of the to-be-processed data in transmission through the memory of the compute node, breaks through a bottleneck of memory delay and bandwidth, and increases a co-processing speed of the to-be-processed data.

Example

Embodiment 2

[0069]An embodiment of the present invention provides a co-processing acceleration method, which is used to increase a speed of co-processing in a computer system. As shown in FIG. 3, the method includes:

[0070]S201: Receive at least one co-processing request message sent by a compute node in a computer system.

[0071]In an embodiment, each co-processing message carries address information of to-be-processed data (that is, to-be-processed data on which processing is requested by the by the compute node through the co-processing message) corresponding to the co-processing message.

[0072]Specifically, in the computer system, at least one compute node and at least one co-processor card exist. The co-processor card may aid the compute node in task processing, that is, co-processing. When the compute node needs aid of the co-processor card in task processing, the compute node sends a co-processing request message. In an embodiment, the co-processing request message may be a data ...

Example

Embodiment 3

[0109]The embodiment of the present invention provides a co-processing task management apparatus, configured to manage co-processing tasks in a computer system in a unified manner. As shown in FIG. 4, the co-processing task management apparatus includes:

[0110]A message receiving module 420 is configured to receive at least one co-processing request message sent by a compute node in the computer system, where the co-processing request message carries address information of to-be-processed data.

[0111]Specifically, in the computer system, if the compute node needs a co-processor card to process the to-be-processed data, the compute node sends a co-processing request message to the message receiving module 420. The message receiving module 420 receives the co-processing request message sent by the compute node. Content included in the co-processing request message are exactly the same as content of the co-processing request message described in S101 of Embodiment 1 of the pr...

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

An embodiment of the present invention discloses a co-processing acceleration method, including: receiving a co-processing request message which is sent by a compute node in a computer system and carries address information of to-be-processed data; according to the co-processing request message, obtaining the to-be-processed data, and storing the to-be-processed data in a public buffer card; and allocating the to-be-processed data stored in the public buffer card to an idle co-processor card in the computer system for processing. An added public buffer card is used as a public data buffer channel between a hard disk and each co-processor card of a computer system, and to-be-processed data does not need to be transferred by a memory of the compute node, which avoids overheads of the data in transmission through the memory of the compute node, and thereby breaks through a bottleneck of memory delay and bandwidth, and increases a co-processing speed.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application is a continuation of International Application No. PCT / CN2011 / 083770, filed on Dec. 9, 2011, which are hereby incorporated by reference in their entireties.FIELD OF THE INVENTION[0002]The present invention relates to the computer field, and in particular, to a co-processing acceleration method, an apparatus, and a system.BACKGROUND OF THE INVENTION[0003]With the development of computer technologies, computers are applied in increasingly wider fields. In addition to common computer office applications in everyday life, the computers are also applied in some very complex fields, such as large-scale scientific computing and massive data processing, which usually have higher requirements on the processing capability of the computers. However, the processing capability of a single computer is limited, and is likely to become a bottleneck of improving system performance in the foregoing large-scale computing scenarios, and this...

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): G06F13/36
CPCG06F15/167G06F9/3881G06F9/544G06F9/505
Inventor ZHANG, XIAOFENGFANG, FANQIN, LING
Owner HUAWEI TECH CO LTD
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap