Abstract

Communication among nodes plays an important role in any system's throughput. In some architecture, communication plays a bottleneck which degrades the computation power. However, making a high-performance computational system along with the energy-efficient system is incompatible in many cases. In this paper, we present a routing methodology in an optical networks on chip (O-NoC). The proposed methodology includes architectural design as well as algorithms working on that architecture. These algorithms are used to communicate among nodes, which eliminates path reservation criteria and provides guaranteed communication. Under some criteria, our methodology achieved excellent results in reducing design space complexity. Furthermore, we evaluated our results on synthetic traffic patterns which include random uniform and hotspot traffic patterns. The proposed methodology achieves better results in case of reduction in hop count between nodes, 100 % link utilization which eliminates ideal node and reduces algorithm running time that makes architecture more energy efficient. These achieved results may enhance computation power and communication power at the same time in O-NoC.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call