Forecasting-Based Dynamic Virtual Channels Allocation for Power Optimization of Network-on-Chips

Abstract

In this paper, we present a dynamic power management technique for optimizing the use of virtual channels in network on chips. The technique which is called dynamic virtual channels allocation (DVCA) makes use of the traffic conditions and past buffer utilization to dynamically forecast the number of virtual channels that should be active. In this technique, for low(high) traffic loads, a small (large) number of VCs are allocated to the corresponding input channel. This provides us with the ability to reduce the power consumption of the router while maintaining the data communication rate. To assess the efficacy of the proposed method, the network on chip has been simulated using several traffic profiles. The simulation results show that up to 35% reduction in the buffer power consumption and up to 20% savings in the overall router power consumption may be achieved. Finally, the area and power overheads of the technique are negligible.