Fully Row/Column-Parallel In-memory Computing SRAM Macro employing Capacitor-based Mixed-signal Computation with 5-b Inputs

Jinseok Lee, Hossein Valavi, Yinqi Tang, Naveen Verma

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

1 Scopus citations

Abstract

This paper presents an in-memory computing (IMC) macro in 28nm for fully row/column-parallel matrix-vector multiplication (MVM), exploiting precise capacitor-based analog computation to extend from binary input-vector elements to 5-b input-vector elements, for 16x increase in energy efficiency and 5x increase in throughput. The 1152(row)x256(col.) macro employs multilevel input drivers based on a digital-switch DAC implementation, which preserve compute accuracy well beyond the 8-b resolution of the output ADCs, and whose area is halved via a dynamic-range doubling (DRD) technique. The macro achieves the highest reported IMC energy efficiency of 5796 TOPS/W and compute density of 12 TOPS/mm2 (both normalized to 1-b ops). CIFAR-10 image classification is demonstrated with accuracy of 91%, equal to the level of ideal SW implementation.

Original languageEnglish (US)
Title of host publication2021 Symposium on VLSI Technology, VLSI Technology 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9784863487802
StatePublished - 2021
Event41st Symposium on VLSI Technology, VLSI Technology 2021 - Virtual, Online, Japan
Duration: Jun 13 2021Jun 19 2021

Publication series

NameDigest of Technical Papers - Symposium on VLSI Technology
Volume2021-June
ISSN (Print)0743-1562

Conference

Conference41st Symposium on VLSI Technology, VLSI Technology 2021
Country/TerritoryJapan
CityVirtual, Online
Period6/13/216/19/21

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Fully Row/Column-Parallel In-memory Computing SRAM Macro employing Capacitor-based Mixed-signal Computation with 5-b Inputs'. Together they form a unique fingerprint.

Cite this