Collective Communication in Hexagonal Mesh and King Networks On-Chip

Project: Other project

Project Details

Description

Network on chip is an emerging paradigm for communications within large VLSI systems implemented on a single silicon chip. A NoC is constructed from multiple point-to-point data links interconnected by switches (a.k.a. routers), such that messages can be relayed from any source module to any destination module over several links, by making routing decisions at the switches. In this research work we will develop new collective communication algorithms for hexagonal Mesh and King network on-chip (NOC). Hexagonal mesh and King networks have gained popularity as interconnection networks because of their very good topological properties. Among the collective communication algorithms we focus on the one-to-all broadcast where a source node send a message to all the node in the NOC and the all-to-all broadcast where every node is send a message to all other node in the NOC. The new proposed communication algorithms use spanning trees in order to minimize the delay time. We also address the issue of reducing the buffer size in the routers which in turns minimizes the power consumption. We will conduct analytical and simulation based performance evaluation to measure the performance of the proposed communication algorithms in terms of power consumption and communication delays when transferring data in NoC architectures.

Layman's description

Network on chip is an emerging paradigm for communications within large VLSI systems implemented on a single silicon chip. A NoC is constructed from multiple point-to-point data links interconnected by switches (a.k.a. routers), such that messages can be relayed from any source module to any destination module over several links, by making routing decisions at the switches. In this research work we will develop new collective communication algorithms for hexagonal Mesh and King network on-chip (NOC). Hexagonal mesh and King networks have gained popularity as interconnection networks because of their very good topological properties. Among the collective communication algorithms we focus on the one-to-all broadcast where a source node send a message to all the node in the NOC and the all-to-all broadcast where every node is send a message to all other node in the NOC. The new proposed communication algorithms use spanning trees in order to minimize the delay time. We also address the issue of reducing the buffer size in the routers which in turns minimizes the power consumption. We will conduct analytical and simulation based performance evaluation to measure the performance of the proposed communication algorithms in terms of power consumption and communication delays when transferring data in NoC architectures.
AcronymTTotP
StatusNot started

Keywords

  • Collective

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.