Skip to main navigation Skip to search Skip to main content

Open-source asynchronous periphery for mixed-signal neuromorphic chips

  • Hugh Greatorex
  • , Madison Cotteret
  • , Michele Mastella
  • , Arianna Rubino
  • , Elisabetta Chicca
  • , Ole Richter
  • University of Groningen
  • Neuronova
  • Swiss Federal Institute of Technology Zurich

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

1 Downloads (Orbit)

Abstract

Mixed-signal neuromorphic systems combine analog neural dynamics with event-driven asynchronous digital communication. Designing the asynchronous digital periphery for these systems is often a major practical barrier however, in part due to the lack of available reusable, silicon-verified components. This forces many research groups to re-design core components from scratch, increasing development time and risk. To address this, we present an open-source library of silicon-verified asynchronous peripheral circuits designed for mixed-signal neuromorphic processors, using the Asynchronous Circuit Toolkit (ACT) with Quasi-Delay-Insensitive (QDI) design methodology and Dual-Rail (DR) encoding. We describe the circuit templates, outline a multi-level verification methodology from behavioural simulation through post-fabrication testing, and present timing measurements from fabricated silicon.
Original languageEnglish
Title of host publicationProceedings of the International Conference on Neuromorphic Systems, ICONS'26
PublisherAssociation for Computing Machinery
Publication date2026
Pages67-74
ISBN (Electronic)79-8-4007-2808-2/26/08
DOIs
Publication statusPublished - 2026
EventInternational Conference on Neuromorphic Systems 2026 - Chicago, United States
Duration: 4 Aug 20266 Aug 2026

Conference

ConferenceInternational Conference on Neuromorphic Systems 2026
Country/TerritoryUnited States
CityChicago
Period04/08/202606/08/2026

Fingerprint

Dive into the research topics of 'Open-source asynchronous periphery for mixed-signal neuromorphic chips'. Together they form a unique fingerprint.

Cite this