A greedy approach for latency-bounded deadlock-free routing path allocation for application-specific NoCs

Amit Verma, Pritpal S. Multani, Daniel Mueller-Gritschneder, Vladimir Todorov, Ulf Schlichtmann

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

8 Scopus citations

Abstract

Custom network-on-chip (NoC) structures have improved power and area metrics compared to regular NoC topologies for application-specific systems-on-a-chip (SoCs). The synthesis of an application-specific NoC is a combinatorial problem. This paper presents a novel heuristic for solving the routing path allocation step. Its main advantages are the support of realistic nonlinear cost estimation and the ability to handle latency constraints, which guarantee high performance of processing elements sensitive to communication delays. Additionally, the method generates deadlock-free routing by avoiding cycles in the channel dependency graph. The NoC is constructed sequentially in a greedy manner by selecting the routing path for each communication flow in such a way that the additional NoC HW resources are kept minimal. The routing path is found using a binary search cheapest bounded path (BSCBP) algorithm. The method is highly efficient and provides a NoC routing path allocation for a smart phone SoC with 25 processing elements and 96 flows in less than a minute.

Original languageEnglish
Title of host publication2013 7th IEEE/ACM International Symposium on Networks-on-Chip, NoCS 2013
DOIs
StatePublished - 2013
Event2013 7th IEEE/ACM International Symposium on Networks-on-Chip, NoCS 2013 - Tempe, AZ, United States
Duration: 21 Apr 201324 Apr 2013

Publication series

Name2013 7th IEEE/ACM International Symposium on Networks-on-Chip, NoCS 2013

Conference

Conference2013 7th IEEE/ACM International Symposium on Networks-on-Chip, NoCS 2013
Country/TerritoryUnited States
CityTempe, AZ
Period21/04/1324/04/13

Fingerprint

Dive into the research topics of 'A greedy approach for latency-bounded deadlock-free routing path allocation for application-specific NoCs'. Together they form a unique fingerprint.

Cite this