Abstract

Natural organic memristors manufactured using honey have exhibited promising synaptic behavior, offering notable advantages such as environmental sustainability, low production and disposal costs, non-volatile storage capabilities, and bio/CMOS compatibility. In this paper, experimental evaluations of honey-memristor based neuromorphic systems are reported with a focus on estimating their carbon footprint. First, manufacturing and testing of honey-memristors is briefly described. To suppress variations and improve accuracy, an optimization technique - parallel memristors is applied. Experimental results indicate that this optimization method significantly improves the device performance of up to. However, it also leads to higher energy consumption and a higher carbon footprint of ≈ 0.44 gCO2, particularly when the honey-memristor has more conductance levels. Also, experimental results indicate that the performance of 128 - level honey-memristor based neuromorphic system is the best amongst all devices to balance performance and carbon emission. This finding underscores an important trade-off between enhanced device performance and environmental sustainability, highlighting the need for developing energy-efficient algorithms. To the best of our knowledge, this is the first study to estimate the carbon footprint of a natural organic memristor-based neuromorphic system.

Department(s)

Electrical and Computer Engineering

Publication Status

Open Access

Comments

U.S. Department of Energy, Grant 2247342

Keywords and Phrases

Emerging non-volatile devices; Memristor; Neuromorphic Computing

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2025 Association for Computing Machinery, All rights reserved.

Publication Date

29 Jun 2025

Download

Full Text Link

Share

 
COinS