Distilling Arbitration Logic from Traces using Machine Learning: A Case Study on NoC

Arbitration logic is extensively used in modern computer architectures to dynamically determine how shared hardware resources are allocated or accessed. Recent work has shown that machine learning techniques can learn non-obvious yet effective arbitration policies, which in simulation demonstrate su...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2021 58th ACM/IEEE Design Automation Conference (DAC) s. 55 - 60
Hlavní autoři: Zhou, Yuan, Wang, Hanyu, Yin, Jieming, Zhang, Zhiru
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 05.12.2021
Témata:
Computational modeling
Deep learning
Machine Learning
Network-on-Chip
Neural networks
Ports (computers)
Throughput
Traffic control
Training
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Arbitration logic is extensively used in modern computer architectures to dynamically determine how shared hardware resources are allocated or accessed. Recent work has shown that machine learning techniques can learn non-obvious yet effective arbitration policies, which in simulation demonstrate superior performance over human-designed heuristics. However, existing methods based on deep learning are too expensive to be directly implemented as an arbitration unit in hardware. While some prior efforts managed to manually analyze and reduce a deep learning model into relatively small circuits in certain cases, such ad hoc and labor-intensive approaches cannot easily generalize. In this work, we propose a new methodology to automatically "distill" the arbitration logic from simulation traces. Starting by training a deep learning model, we leverage tree-based models as a bridge to convert the more complex model to a compact logic implementation. This paper presents a case study of the proposed methodology on a network-on-chip port arbitration task. Compared with an array of combinational multipliers that exactly computes the neural network output, our arbitration logic achieves up to 282x area reduction without significant performance degradation. Under the training traffic, our arbitration logic achieves up to 64x reduction in average packet latency and up to 5% increase in network throughput over the FIFO arbitration policy. The distilled arbitration policy is also able to generalize to different injection rates and traffic patterns.
DOI:10.1109/DAC18074.2021.9586301