AdEle: An Adaptive Congestion-and-Energy-Aware Elevator Selection for Partially Connected 3D NoCs

TL;DR

This paper introduces AdEle, an adaptive elevator-selection scheme for partially connected 3D NoCs that optimizes traffic distribution, reducing latency and energy consumption through a combination of offline and online algorithms.

Contribution

It presents a novel adaptive congestion- and energy-aware elevator-selection scheme using multi-objective simulated annealing and online policy for PC-3DNoCs.

Findings

Reduces network latency by up to 14.6%.

Achieves 10.9% average latency improvement.

Maintains less than 6.9% energy overhead.

Abstract

By lowering the number of vertical connections in fully connected 3D networks-on-chip (NoCs), partially connected 3D NoCs (PC-3DNoCs) help alleviate reliability and fabrication issues. This paper proposes a novel, adaptive congestion- and energy-aware elevator-selection scheme called AdEle to improve the traffic distribution in PC-3DNoCs. AdEle employs an offline multi-objective simulated-annealing-based algorithm to find good elevator subsets and an online elevator selection policy to enhance elevator selection during routing. Compared to the state-of- the-art techniques under different real-application traffics and configuration scenarios, AdEle improves the network latency by 10.9% on average (up to 14.6%) with less than 6.9% energy consumption overhead.