Predicting Networks-on-Chip traffic congestion with Spiking Neural Networks

Network congestion is one of the critical reasons for degradation data throughput performance in Networks-on-Chip (NoCs), with delays caused by data-buffer queuing routers. Local buffer or router impacts on network as it gradually spreads to neighbouring routers and beyond. In this paper, we propose a novel approach NoC traffic prediction using Spiking Neural Networks (SNNs) focus predicting local so minimize its impact overall NoCs throughput. The key novelty utilizing SNNs recognize temporal patterns from buffers hotspots. We investigate two neural models, Leaky Integrate Fire (LIF) Spike Response Model (SRM) check terms coverage. Results accuracy precision are reported synthetic real-time multimedia applications simulation results LIF based predictor providing an average 88.28%–96.25% 82.09%–96.73% compared 85.25%–95.69% 73% 98.48% SRM model when looking at formations 30 clock cycles advance actual hotspot occurrence.