Logo image
Back
Reliability-Performance Trade-offs in Neuromorphic Computing
Conference proceeding   Open access

Reliability-Performance Trade-offs in Neuromorphic Computing

Twisha Titirsha and Anup Das
2020 11th International Green and Sustainable Computing Workshops (IGSC), pp 1-5
19 Oct 2020
DOI: https://doi.org/10.1109/IGSC51522.2020.9290845
url
http://arxiv.org/abs/2009.12672View
SubmittedOpen Access (License Unspecified),  Open

Abstract

cost-per-bit crossbar Endurance energy efficiency Hardware learning (artificial intelligence) longer current paths lower endurance Machine learning machine learning tasks Memory management neural nets Neural Networks neuromorphic architectures neuromorphic computing Non-Volatile Memory (NVM) NonVolatile Memory NVM cells parasitic components parasitic voltage Phase change materials Phase-Change Memory (PCM) power aware computing random-access storage Reliability reliability-performance trade-offs shorter current paths significant asymmetry SNN mapping technique Task analysis voltage drop
Neuromorphic architectures built with Non-Volatile Memory (NVM) can significantly improve the energy efficiency of machine learning tasks designed with Spiking Neural Networks (SNNs). A major source of voltage drop in a crossbar of these architectures are the parasitic components on the crossbar's bitlines and wordlines, which are deliberately made longer to achieve lower cost-per-bit. We observe that the parasitic voltage drops create a significant asymmetry in programming speed and reliability of NVM cells in a crossbar. Specifically, NVM cells that are on shorter current paths are faster to program but have lower endurance than those on longer current paths, and vice versa. This asymmetry in neuromorphic architectures create reliability-performance trade-offs, which can be exploited efficiently using SNN mapping techniques. In this work, we demonstrate such trade-offs using a previously-proposed SNN mapping technique with 10 workloads from contemporary machine learning tasks for a state-of-the art neuromoorphic hardware.

Metrics

11 Record Views
15 citations in Scopus

Details

Logo image