A novel analytical model for wormhole switching network on chip with adaptive routing

This paper proposes a novel analytical model for predicting average message latency under different traffic load in a network-on-chip where wormhole switching and fully adaptive routing protocols are used. The model has simple close-form calculations and yields extremely accurate results in the network stable regions. The validity of this model is shown by comparing analytical results with simulation results obtained on high-radix k-ary 2-cube networks.