AR-CIM: A 460.22TOPS/W SRAM-based Analog Reconfigurable Computing-in-Memory Macro with 1/2/4/8-Bit Variable Precision

Byeongseon Choi, Edward Choi, Donghyeon Yi, Jiho Chun, Sohmyung Ha, Ik Joon Chang, Minkyu Je

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

Abstract

This work presents an SRAM-based analog reconfigurable computing-in-memory macro with 409.6-GOPS throughput and 460.22-TOPS/W energy efficiency. The proposed reconfigurable macro supports 1/2/4/8b inputs and 1/2/4/8b signed weights while maintaining 100% memory utilization. Furthermore, we propose (i) a dual-domain DAC to achieve high-precision input with low area overhead and (ii) an input-sense-and-skip SAR ADC, which reduces the average power overhead by 54% compared with conventional SAR ADC.

Original languageEnglish (US)
Title of host publicationESSERC 2024 - Proceedings
Subtitle of host publication50th IEEE European Solid-State Electronics Research Conference
PublisherIEEE Computer Society
Pages361-364
Number of pages4
ISBN (Electronic)9798350388138
DOIs
StatePublished - 2024
Event50th IEEE European Solid-State Electronics Research Conference, ESSERC 2024 - Bruges, Belgium
Duration: Sep 9 2024Sep 12 2024

Publication series

NameEuropean Solid-State Circuits Conference
ISSN (Print)1930-8833

Conference

Conference50th IEEE European Solid-State Electronics Research Conference, ESSERC 2024
Country/TerritoryBelgium
CityBruges
Period9/9/249/12/24

Keywords

  • Analog Computing-in-memory (ACIM)
  • Convolution Neural Network
  • SRAM

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'AR-CIM: A 460.22TOPS/W SRAM-based Analog Reconfigurable Computing-in-Memory Macro with 1/2/4/8-Bit Variable Precision'. Together they form a unique fingerprint.

Cite this