SupeRBNN: Randomized Binary Neural Network Using Adiabatic Superconductor Josephson Devices
Preprint
Abstract
Adiabatic Quantum-Flux-Parametron (AQFP) is a superconducting logic with extremely high energy efficiency. By employing the distinct polarity of current to denote logic `0' and `1', AQFP devices serve as excellent carriers for binary neural network (BNN) computations. Although recent research has made initial strides toward developing an AQFP-based BNN accelerator, several critical challenges remain, preventing the design from being a comprehensive solution. In this paper, we propose SupeRBNN, an AQFP-based randomized BNN acceleration framework that leverages software-hardware co-optimization to eventually make the AQFP devices a feasible solution for BNN acceleration. Specifically, we investigate the randomized behavior of the AQFP devices and analyze the impact of crossbar size on current attenuation, subsequently formulating the current amplitude into the values suitable for use in BNN computation. To tackle the accumulation problem and improve overall hardware performance, we propose a stochastic computing-based accumulation module and a clocking scheme adjustment-based circuit optimization method. We validate our SupeRBNN framework across various datasets and network architectures, comparing it with implementations based on different technologies, including CMOS, ReRAM, and superconducting RSFQ/ERSFQ. Experimental results demonstrate that our design achieves an energy efficiency of approximately 7.8x10^4 times higher than that of the ReRAM-based BNN framework while maintaining a similar level of model accuracy. Furthermore, when compared with superconductor-based counterparts, our framework demonstrates at least two orders of magnitude higher energy efficiency.Keywords
All Related Versions
This publication has 34 references indexed in Scilit:
- Cryogenic memory technologiesNature Electronics, 2023
- ReActNet: Towards Precise Binary Neural Network with Generalized Activation Functions2020
- Low-latency power-dividing clocking scheme for adiabatic quantum-flux-parametron logicApplied Physics Letters, 2020
- Adiabatic Quantum-Flux-Parametron: Towards Building Extremely Energy-Efficient Circuits and SystemsScientific Reports, 2019
- A reversible full adder using adiabatic superconductor logicSuperconductor Science and Technology, 2019
- An adiabatic superconductor 8-bit adder with 24kBT energy dissipation per junctionApplied Physics Letters, 2019
- XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks2016
- Adiabatic quantum-flux-parametron cell library adopting minimalist designJournal of Applied Physics, 2015
- Measurement of 10 zJ energy dissipation of adiabatic quantum-flux-parametron logic using a superconducting resonatorApplied Physics Letters, 2013
- An adiabatic quantum flux parametron as an ultra-low-power logic deviceSuperconductor Science and Technology, 2013