Q-StaR: A Quasi-Static Routing Scheme for NoCs

TL;DR

Q-StaR introduces a quasi-static routing scheme for NoCs that predicts load distribution trends to improve load balancing, significantly enhancing throughput and reducing latency compared to traditional static routing.

Contribution

It identifies key factors influencing load distribution and develops N-Rank and BiDOR to adapt routing decisions dynamically while maintaining simplicity.

Findings

42.9% throughput improvement under uniform traffic

86.4% reduction in mean latency under realistic workloads

95.3% reduction in maximum latency

Abstract

In networks-on-chip, static routing schemes are favored for their simplicity and predictability, but they cannot effectively balance network load due to the unawareness of runtime load distribution. Q-StaR discovers two factors (topology and traffic distribution) that determine the long-term trend of load distribution, and proposes N-Rank to extract this trend. The obtained information is used to guide BiDOR's route selection at runtime, thereby improving load balancing while retaining simplicity and predictability. Simulation validates that Q-StaR significantly outperforms the typical dimension-order routing (throughput under uniform traffic improved by 42.9\%, and mean/maximum latency under realistic workloads reduced by 86.4\%/95.3\%).