TNNGen: Automated Design of Neuromorphic Sensory Processing Units for Time-Series Clustering
This work addresses the labor-intensive process of designing neuromorphic hardware for time-series clustering, offering a domain-specific automation tool for researchers and engineers.
The paper tackles the problem of manual hardware design for Temporal Neural Networks (TNNs) by introducing TNNGen, an automated tool that converts PyTorch models to post-layout netlists, reducing design runtimes and enabling accurate silicon metric forecasting without full hardware process flow.
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.