Fabrication of Efficient Reconfigurable Neuromorphic Systems

Abstract

The goal of this effort was to fabricate hybrid CMOS/RRAM chips that could be used for a variety of Air Force specific applications, using a low-power, reconfigurable framework. Working with collaborators at UT-Knoxville and a number of other institutions, the SUNY Poly team was able to compile a design containing an unique neuromorphic, reconfigurable test chip (UT-Knoxville design), a variety of hybrid CMOS/RRAM demonstration circuits (from UT-Austin, ASU), and a neuromorphic computing demonstrator chip (from UT-San Antonio). In addition, the SUNY Poly team designed multiple memory arrays for demonstration of compute in memory operations and analog information encoding. To this end, the SUNY Poly team also developed optimized RRAM testing parameters to 1) maintain high device yield and endurance, 2) enable a large device memory window, and 3) perform analog information encoding using sub-nanosecond pulses. Together, this work provides a framework for ongoing testing and evaluation of hybrid CMOS/RRAM circuits by the SUNY Poly team and collaborators. Full testing of the circuits and chips fabricated under this effort is underway at the partnering institutions and will continue under the OUSD "NeuroPipe" ARAP program.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2023
Accession Number
AD1190520

Entities

People

  • Nathaniel C Cady

Tags

Communities of Interest

  • Advanced Electronics

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Artificial Intelligence
  • Artificial Intelligence Software
  • Complementary Metal-Oxide Semiconductors
  • Computational Science
  • Computer Architecture
  • Computer Programming
  • Computers
  • Detection
  • Electronics Industry
  • Electronics Laboratories
  • Energy Consumption
  • Field Effect Transistors
  • Graphical User Interface
  • Information Exchange
  • Information Science
  • Machine Learning
  • Materials
  • Network Science
  • Neural Networks
  • Physical Vapor Deposition
  • Semiconductors
  • Unmanned Vehicles

Readers

  • Distributed Systems and Data Platform Development
  • Integrated Circuit Design and Technology.