Processing-In-Memory Acceleration of Convolutional Neural Networks for Energy-Effciency, and Power-Intermittency Resilience

Arman Roohi, Shaahin Angizi, Deliang Fan, Ronald F. Demara

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Scopus citations

Abstract

Herein, a bit-wise Convolutional Neural Network (CNN) in-memory accelerator is implemented using Spin-Orbit Torque Magnetic Random Access Memory (SOT-MRAM) computational sub-arrays. It utilizes a novel AND-Accumulation method capable of significantly-reduced energy consumption within convolutional layers and performs various low bitwidth CNN inference operations entirely within MRAM. Power-intermittence resiliency is also enhanced by retaining the partial state information needed to maintain computational forward-progress, which is advantageous for battery-less IoT nodes. Simulation results indicate ∼ 5.4× higher energy-efficiency and 9× speedup over ReRAM-based acceleration, or roughly ∼ 9.7× higher energy-efficiency and 13.5× speedup over recent CMOS-only approaches, while maintaining inference accuracy comparable to baseline designs.

Original languageEnglish (US)
Title of host publicationProceedings of the 20th International Symposium on Quality Electronic Design, ISQED 2019
PublisherIEEE Computer Society
Pages8-13
Number of pages6
ISBN (Electronic)9781728103921
DOIs
StatePublished - Apr 23 2019
Externally publishedYes
Event20th International Symposium on Quality Electronic Design, ISQED 2019 - Santa Clara, United States
Duration: Mar 6 2019Mar 7 2019

Publication series

NameProceedings - International Symposium on Quality Electronic Design, ISQED
Volume2019-March
ISSN (Print)1948-3287
ISSN (Electronic)1948-3295

Conference

Conference20th International Symposium on Quality Electronic Design, ISQED 2019
Country/TerritoryUnited States
CitySanta Clara
Period3/6/193/7/19

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering
  • Safety, Risk, Reliability and Quality

Fingerprint

Dive into the research topics of 'Processing-In-Memory Acceleration of Convolutional Neural Networks for Energy-Effciency, and Power-Intermittency Resilience'. Together they form a unique fingerprint.

Cite this