TNNGen: Automated Design of Neuromorphic Sensory Processing Units for Time-Series Clustering

TL;DR

TNNGen is an automated framework that converts PyTorch models of Temporal Neural Networks into hardware designs, enabling efficient and accurate exploration of neuromorphic sensory processing units for time-series clustering.

Contribution

It introduces the first open-source, automated toolchain for designing TNN hardware from software models, reducing manual effort and enabling silicon metric prediction.

Findings

Successfully simulated seven TNN designs for time-series clustering.

Demonstrated accurate prediction of hardware complexity and silicon metrics.

Showcased the framework's ability to forecast hardware performance without physical fabrication.

Abstract

Temporal Neural Networks (TNNs), a special class of spiking neural networks, draw inspiration from the neocortex in utilizing spike-timings for information processing. Recent works proposed a microarchitecture framework and custom macro suite for designing highly energy-efficient application-specific TNNs. These recent works rely on manual hardware design, a labor-intensive and time-consuming process. Further, there is no open-source functional simulation framework for TNNs. This paper introduces TNNGen, a pioneering effort towards the automated design of TNNs from PyTorch software models to post-layout netlists. TNNGen comprises a novel PyTorch functional simulator (for TNN modeling and application exploration) coupled with a Python-based hardware generator (for PyTorch-to-RTL and RTL-to-Layout conversions). Seven representative TNN designs for time-series signal clustering across diverse sensory modalities are simulated and their post-layout hardware complexity and design runtimes are assessed to demonstrate the effectiveness of TNNGen. We also highlight TNNGen's ability to accurately forecast silicon metrics without running hardware process flow.