Determining Effective Shortest Path in Asynchronous Network-on-Chip Through Bio-Inspired Optimization Techniques

Abstract

An asynchronous network-on-chip (ANoC) provides several benefits over a clocked network present in a system-on-chips (SoC). A few published techniques are available for aiding in the automation of high-level ANoC design for the case of fixed-function SoC. An effort is made to simulate an ANoC with mesh topology coupled with hopfield neural network. Bio-inspired algorithms are employed in the ANoC to determine the shortest path through which data packets are transferred. This shortest path search is carried out to reduce the operational complex issues. Particle swarm optimization, ant colony optimization, firefly optimization algorithm and cuckoo search algorithm are used in this work to find the shortest path. From the simulations, vital parameters like throughput, energy, latency are extracted and are compared to identify the efficient technique to generate the shortest path in the ANoC.