Abstract: This work presents an approach to optimize the network and reduce the congestion within a chip. It involves using NVIDIA Simba architecture as reference for Processing Element (PE).