Traffic-Balanced IP Mapping Algorithm for 2D-mesh on-chip-networks

Ting Jung Lin, Shu Yen Lin, An Yeu Wu

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

3 Citations (Scopus)


Intellectual Properties (IPs) mapping algorithms for On-Chip-Networks (OCNs) allocate a set of IPs onto given network topologies. The existing mapping algorithms limit a single IP to connect to a single router. Hence, the IPs with large communication volumes will result in heavy traffic loads of certain routers. Those routers may become hot spots due to high power density, which affects the reliability of chips. In this paper, new Network Interfaces (NIs) were proposed to solve the aforementioned problem. Traffic-Balanced Mapping Algorithm (TBMAP) is also proposed based on the new NIs. The traffic loads then become more decentralized, and the traffic of all the routers on the chip can be balanced without sacrificing the networking performance. The TBMAP has short runtime to achieve balanced network traffic loads, which leads to the enhanced performance of OCNs. The experimental results show that at least 24% communication time is saved for real applications.

Original languageEnglish
Title of host publication2008 IEEE Workshop on Signal Processing Systems, SiPS 2008, Proceedings
Number of pages4
Publication statusPublished - 2008
Externally publishedYes
Event2008 IEEE Workshop on Signal Processing Systems, SiPS 2008 - Washington, DC, United States
Duration: 8 Oct 200810 Oct 2008

Publication series

NameIEEE Workshop on Signal Processing Systems, SiPS: Design and Implementation
ISSN (Print)1520-6130


Conference2008 IEEE Workshop on Signal Processing Systems, SiPS 2008
Country/TerritoryUnited States
CityWashington, DC


  • Interconnection networks
  • Network interfaces

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing
  • Applied Mathematics
  • Hardware and Architecture


Dive into the research topics of 'Traffic-Balanced IP Mapping Algorithm for 2D-mesh on-chip-networks'. Together they form a unique fingerprint.

Cite this