TY  - CHAPTER

JF  - Biologically-Inspired Collaborative Computing

T1  - Self-optimized Routing in a Network on-a-Chip

SP  - 199

EP  - 212

PY  - 2008///

UR  - http://dx.doi.org/10.1007/978-0-387-09655-1_18

M3  - 10.1007/978-0-387-09655-1_18

AU  - Trumler, Wolfgang

AU  - Schlingmann, Sebastian

AU  - Ungerer, Theo

AU  - Bahn, Jun

AU  - Bagherzadeh, Nader

N2  - Many-cores are on the cusp of becoming state-of-the-art processor technology for the next decade. To guarantee efficient communication

 between multiple cores, a Network-on-a-Chip (NoC) is considered as an alternative to overcome the limitations of the ubiquitous

 bus technology.

 

 In this paper, we present an approach to further improve the routing in an NoC with a self-optimized routing strategy. We

 extended the routers of a network to measure their load and to send an appropriate load information to their direct neighbors.

 The load information is used to decide in which direction a packet should be routed to avoid hot-spots. Evaluation results

 show a significant increase in the network throughput. With the self-optimized routing, the NoC is capable of routing up to

 two times more packets compared to the original routing algorithm proposed by Lee and Bagherzadeh, 2006.

 

 

 

ER  -