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

A deadlock-free adaptive routing algorithm in a Torus network

A deadlock-free, self-adaptive technology, applied in the field of distributed networks, can solve problems such as network delay reduction, achieve the effect of avoiding deadlocks and improving data transmission efficiency

Active Publication Date: 2013-07-10
TSINGHUA UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The virtual pass-through method requires each node to have a large data cache area, but compared with store-and-forward, the network delay is greatly reduced

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
  • A deadlock-free adaptive routing algorithm in a Torus network
  • A deadlock-free adaptive routing algorithm in a Torus network
  • A deadlock-free adaptive routing algorithm in a Torus network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0045] The Torus network adopts the virtual cut-through switching method without flow control. The algorithm is implemented as follows:

[0046] 1) The R1 channel is fully adaptive, and data packets can request the R1 channel at any time;

[0047] 2) If the data packet is already in the Mesh subnet, that is, the data packet no longer needs to go through any turnaround link to reach the destination, the data packet can request the R2 channel, and the data packet should follow the Mesh network when using the R2 channel Deadlock-free routing algorithms in , such as dimension order algorithm, turn model (TurnModel), etc.;

[0048] 3) If the next hop of the data packet is through a turnaround link, and the dimension of the turnaround link is the lowest dimension among the turnaround links that the data packet needs to pass through to reach the destination, then on that turnaround link , the packet can request the R2 channel.

[0049] For a Torus network with n dimensions and k no...

Embodiment 2

[0099] The Torus network adopts the virtual cut-through switching mode and uses flow control, so it is necessary to first define the secure data packet and the non-secure data packet:

[0100] 1) If the data packet is already in the Mesh subnet, that is, the data packet does not need to go through any turnaround link to reach the destination, and the next hop of the data packet follows the deadlock-free routing algorithm in the Mesh network, then for the next The node on one hop, the data packet is a security data packet;

[0101] 2) If the next hop of the data packet is through a turnaround link, and the dimension of the turnaround link is the lowest dimension in the turnaround link that the data packet needs to pass through to reach the destination, for the node on the next hop, The packet is a security packet;

[0102] A data packet that does not meet any of the above conditions is a non-secure data packet,

[0103] After giving the definitions of secure and non-secure da...

Embodiment 3

[0110] If the Torus network adopts the wormhole exchange method, according to the distance from each node to the vertex in the Torus network, all nodes are divided into different sets Si, the set S0 contains all vertices, and the minimum path from the nodes in the set Si to all vertices , the shortest distance is i, the implementation of the algorithm is as follows:

[0111] 1) When the data packet needs to pass through the turnaround link, it can only apply for the R1 channel from the node in the set Si to the node in the set Sj, where j≤i, when the data packet does not need to pass through the turnaround link, the data packet There is no limit to request R1 channel;

[0112] 2) If the data packet is already in the Mesh subnet, that is, the data packet does not need to go through any turnaround link to reach the destination, then the data packet can request the R2 channel, and the data packet should follow the Mesh when using the R2 channel Deadlock-free routing algorithms i...

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 deadlock-free minimal routing algorithm called clue is first disclosed for VCT tori. Only two virtual channels are required. One channel is applied in the deadlock-free routing algorithm for the mesh sub-network based on a known base routing scheme, such as, negative-first or dimension-order routing. The other channel is similar to an adaptive channel. This combination completes a very novel fully-adaptive routing scheme because the first channel does not supply routing paths for every source-destination pair. Based on clue, we proposed two other algorithms named flow controlled clue and wormhole clue. Flow controlled clue is also proposed for VCT-switched tori. It is fully adaptive deadlock-free. Each input port requires at least two buffers, each of which is able to keep a packet. A simple but well-designed flow control function is used in the proposed flow controlled clue routing algorithm to avoid deadlocks. Wormhole clue is proposed for wormhole-switched tori. It is partially adaptive because we add some constraints to the adaptive channel for deadlock avoidance.

Description

technical field [0001] The invention belongs to the technical field of distributed networks, relates to a Torus network, in particular to a deadlock-free self-adaptive routing algorithm in the Torus network. Background technique [0002] Direct Interconnection Network, as a common network topology, has been widely used in multi-processor systems (Multi-processor) and multi-computer systems (Multi-computer). [0003] In a large-scale multi-computer system, each node has its own processor and local memory, and the system realizes the connection between adjacent nodes through a direct network. The biggest advantage of the direct network is that it has good scalability. [0004] In a large-scale multi-computer system, each node has a separate router (Router), which handles communication-related tasks. Although this task can be performed by the corresponding processor, the overlapping execution of calculation and communication tasks in the node can be realized after the router ...

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): H04L12/705H04L45/18
CPCH04L45/18
Inventor 向东罗伟
Owner TSINGHUA UNIV
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