A New Selection Policy for Adaptive Routing in Network on Chip

Efficient and deadlock-free routing is critical to the performance of networks-on-chip. In this paper we present a new selection policy that can be coupled to any adaptive routing algorithm to improve the performance in terms of average delay. The proposed approach introduce the concept of Neighbors-on-Path to exploit the situations of indecision occurring when the routing function returns several admissible output channels. A selection strategy is developed with the aim to choose the channel that will allow the packet to be routed to its destination along a path that is as free as possible of congested nodes. Experimental comparisons between the proposed approach against both the deterministic XY routing and current state-of-the-art of adaptive routing algorithm for different traffic scenarios show an improvement in terms of saturation point respectively of 36% and almost 10% in average. Key–Words: Networks on Chip, adaptive routing, deadlock-free routing