Knowledge-Based Neural Networks for Fast Design Space Exploration of Hybrid Copper-Graphene On-Chip Interconnect Networks

• DOI: 10.1109/temc.2021.3091714
• Publication Date: 2021-07-21T20:13:36Z
• Authors: Rahul Kumar; S. S. Likith Narayan; Somesh Kumar; Sourajeet Roy; Brajesh Kaushik; Ramachandra Achar; Rohit Sharma
• Keywords: Solver; Design space exploration

In this article, an artificial neural network (ANN) is developed in order to predict the per-unit-length (p. u. l.) parameters of hybrid copper-graphene on-chip interconnects from a prior knowledge their structural geometry and layout. The salient feature proposed ANN that it combines p. l. extracted empirical models along with rigorous full-wave electromagnetic solver. As result, referred as knowledge-based (KBNN). KBNN has been found converge same accuracy conventional but at expense far smaller training time costs. much more suitable for performing design space explorations.