Monolithic 3D+-IC based massively parallel compute-in-memory macro for accelerating database and machine learning primitives

Akshay Krishna Ramanathan, Srivatsa Srinivasa Rangachar, Je Min Hung, Chun Ying Lee, Cheng Xin Xue, Sheng Po Huang, Fu Kuo Hsueh, Chang Hong Shen, Jia Min Shieh, Wen Kuan Yeh, Mon Shu Ho, Hariram Thirucherai Govindarajan, Jack Sampson, Meng Fan Chang, Vijaykrishnan Narayanan

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

Abstract

This paper demonstrates the first Monolithic 3D+-IC based Compute-in-Memory (CiM) Macro performing massively parallel beyond-Boolean operations targeting database and machine learning (ML) applications. The proposed CiM technique supports data filtering, sorting, and sparse matrix-matrix multiplication (SpGEMM) operations. Our system exhibits up to 272x speedup and 151x energy savings compared to the ASIC baseline.

Original languageEnglish (US)
Title of host publication2020 IEEE International Electron Devices Meeting, IEDM 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages28.5.1-28.5.4
ISBN (Electronic)9781728188881
DOIs
StatePublished - Dec 12 2020
Event66th Annual IEEE International Electron Devices Meeting, IEDM 2020 - Virtual, San Francisco, United States
Duration: Dec 12 2020Dec 18 2020

Publication series

NameTechnical Digest - International Electron Devices Meeting, IEDM
Volume2020-December
ISSN (Print)0163-1918

Conference

Conference66th Annual IEEE International Electron Devices Meeting, IEDM 2020
CountryUnited States
CityVirtual, San Francisco
Period12/12/2012/18/20

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Electrical and Electronic Engineering
  • Materials Chemistry

Fingerprint Dive into the research topics of 'Monolithic 3D+-IC based massively parallel compute-in-memory macro for accelerating database and machine learning primitives'. Together they form a unique fingerprint.

Cite this