The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications

Joonhyung Kim; Jongsun Park

doi:10.1109/ISOCC56007.2022.10031424

The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications

Joonhyung Kim, Jongsun Park

School of Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Recently, static random-access memory (SRAM) compute-in-memory (CIM) research has been actively studied for energy efficient acceleration of deep neural network (DNN). The SRAM CIM research can be divided into analog CIM (ACIM) and digital CIM (DCIM) depending on the computing mechanism of MAC operation. Although both ACIM and DCIM are claimed energy efficient to accelerate DNN, detailed analysis and comparisons between two CIMs are hard to find. For the CIM designers who need to decide which type of CIMs can be selected for their DNN application, quantitative analysis in terms of energy and accuracy would be greatly helpful. In this paper, we compare the ACIM and DCIM in terms of the energy efficiency (TOPS/W) and accuracies. For ACIM design, BL chargesharing scheme and 3-bit flash ADC are selected for MAC operations, while adder-tree based MAC operations is used for DCIM design. Both approaches are designed with 256x64 macro using 28nm CMOS process. The simulation results show that DCIM improves \times 1.96 of energy efficiency (TOPS/W) and better CIFAR-1O accuracy compared to ACIM approach.

Original language	English
Title of host publication	Proceedings - International SoC Design Conference 2022, ISOCC 2022
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	129-130
Number of pages	2
ISBN (Electronic)	9781665459716
DOIs	https://doi.org/10.1109/ISOCC56007.2022.10031424
Publication status	Published - 2022
Event	19th International System-on-Chip Design Conference, ISOCC 2022 - Gangneung-si, Korea, Republic of Duration: 2022 Oct 19 → 2022 Oct 22

Publication series

Name	Proceedings - International SoC Design Conference 2022, ISOCC 2022

Conference

Conference	19th International System-on-Chip Design Conference, ISOCC 2022
Country/Territory	Korea, Republic of
City	Gangneung-si
Period	22/10/19 → 22/10/22

Bibliographical note

Funding Information:
ACKNOWLEDGMENT This work was supported by the National Research Foundation of Korea grant funded by the Korea government (NRF-2020R1A2C3014820). The EDA tool was supported by the IC Design Education Center(IDEC), Korea.

Publisher Copyright:
© 2022 IEEE.

Keywords

Analog CIM
CIM
Compute-In-Memory
Digital CIM
SRAM

ASJC Scopus subject areas

Artificial Intelligence
Computer Science Applications
Hardware and Architecture
Safety, Risk, Reliability and Quality

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/ISOCC56007.2022.10031424

Cite this

Kim, J., & Park, J. (2022). The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications. In Proceedings - International SoC Design Conference 2022, ISOCC 2022 (pp. 129-130). (Proceedings - International SoC Design Conference 2022, ISOCC 2022). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISOCC56007.2022.10031424

The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications. / Kim, Joonhyung; Park, Jongsun.
Proceedings - International SoC Design Conference 2022, ISOCC 2022. Institute of Electrical and Electronics Engineers Inc., 2022. p. 129-130 (Proceedings - International SoC Design Conference 2022, ISOCC 2022).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Kim, J & Park, J 2022, The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications. in Proceedings - International SoC Design Conference 2022, ISOCC 2022. Proceedings - International SoC Design Conference 2022, ISOCC 2022, Institute of Electrical and Electronics Engineers Inc., pp. 129-130, 19th International System-on-Chip Design Conference, ISOCC 2022, Gangneung-si, Korea, Republic of, 22/10/19. https://doi.org/10.1109/ISOCC56007.2022.10031424

Kim J, Park J. The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications. In Proceedings - International SoC Design Conference 2022, ISOCC 2022. Institute of Electrical and Electronics Engineers Inc. 2022. p. 129-130. (Proceedings - International SoC Design Conference 2022, ISOCC 2022). doi: 10.1109/ISOCC56007.2022.10031424

@inproceedings{d06857cdc1cc4f5f99b2ad3ad8a1ba7d,

title = "The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications",

abstract = "Recently, static random-access memory (SRAM) compute-in-memory (CIM) research has been actively studied for energy efficient acceleration of deep neural network (DNN). The SRAM CIM research can be divided into analog CIM (ACIM) and digital CIM (DCIM) depending on the computing mechanism of MAC operation. Although both ACIM and DCIM are claimed energy efficient to accelerate DNN, detailed analysis and comparisons between two CIMs are hard to find. For the CIM designers who need to decide which type of CIMs can be selected for their DNN application, quantitative analysis in terms of energy and accuracy would be greatly helpful. In this paper, we compare the ACIM and DCIM in terms of the energy efficiency (TOPS/W) and accuracies. For ACIM design, BL chargesharing scheme and 3-bit flash ADC are selected for MAC operations, while adder-tree based MAC operations is used for DCIM design. Both approaches are designed with 256x64 macro using 28nm CMOS process. The simulation results show that DCIM improves \times 1.96 of energy efficiency (TOPS/W) and better CIFAR-1O accuracy compared to ACIM approach.",

keywords = "Analog CIM, CIM, Compute-In-Memory, Digital CIM, SRAM",

author = "Joonhyung Kim and Jongsun Park",

note = "Funding Information: ACKNOWLEDGMENT This work was supported by the National Research Foundation of Korea grant funded by the Korea government (NRF-2020R1A2C3014820). The EDA tool was supported by the IC Design Education Center(IDEC), Korea. Publisher Copyright: {\textcopyright} 2022 IEEE.; 19th International System-on-Chip Design Conference, ISOCC 2022 ; Conference date: 19-10-2022 Through 22-10-2022",

year = "2022",

doi = "10.1109/ISOCC56007.2022.10031424",

language = "English",

series = "Proceedings - International SoC Design Conference 2022, ISOCC 2022",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "129--130",

booktitle = "Proceedings - International SoC Design Conference 2022, ISOCC 2022",

}

TY - GEN

T1 - The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications

AU - Kim, Joonhyung

AU - Park, Jongsun

N1 - Funding Information: ACKNOWLEDGMENT This work was supported by the National Research Foundation of Korea grant funded by the Korea government (NRF-2020R1A2C3014820). The EDA tool was supported by the IC Design Education Center(IDEC), Korea. Publisher Copyright: © 2022 IEEE.

PY - 2022

Y1 - 2022

N2 - Recently, static random-access memory (SRAM) compute-in-memory (CIM) research has been actively studied for energy efficient acceleration of deep neural network (DNN). The SRAM CIM research can be divided into analog CIM (ACIM) and digital CIM (DCIM) depending on the computing mechanism of MAC operation. Although both ACIM and DCIM are claimed energy efficient to accelerate DNN, detailed analysis and comparisons between two CIMs are hard to find. For the CIM designers who need to decide which type of CIMs can be selected for their DNN application, quantitative analysis in terms of energy and accuracy would be greatly helpful. In this paper, we compare the ACIM and DCIM in terms of the energy efficiency (TOPS/W) and accuracies. For ACIM design, BL chargesharing scheme and 3-bit flash ADC are selected for MAC operations, while adder-tree based MAC operations is used for DCIM design. Both approaches are designed with 256x64 macro using 28nm CMOS process. The simulation results show that DCIM improves \times 1.96 of energy efficiency (TOPS/W) and better CIFAR-1O accuracy compared to ACIM approach.

AB - Recently, static random-access memory (SRAM) compute-in-memory (CIM) research has been actively studied for energy efficient acceleration of deep neural network (DNN). The SRAM CIM research can be divided into analog CIM (ACIM) and digital CIM (DCIM) depending on the computing mechanism of MAC operation. Although both ACIM and DCIM are claimed energy efficient to accelerate DNN, detailed analysis and comparisons between two CIMs are hard to find. For the CIM designers who need to decide which type of CIMs can be selected for their DNN application, quantitative analysis in terms of energy and accuracy would be greatly helpful. In this paper, we compare the ACIM and DCIM in terms of the energy efficiency (TOPS/W) and accuracies. For ACIM design, BL chargesharing scheme and 3-bit flash ADC are selected for MAC operations, while adder-tree based MAC operations is used for DCIM design. Both approaches are designed with 256x64 macro using 28nm CMOS process. The simulation results show that DCIM improves \times 1.96 of energy efficiency (TOPS/W) and better CIFAR-1O accuracy compared to ACIM approach.

KW - Analog CIM

KW - CIM

KW - Compute-In-Memory

KW - Digital CIM

KW - SRAM

UR - http://www.scopus.com/inward/record.url?scp=85148440885&partnerID=8YFLogxK

U2 - 10.1109/ISOCC56007.2022.10031424

DO - 10.1109/ISOCC56007.2022.10031424

M3 - Conference contribution

AN - SCOPUS:85148440885

T3 - Proceedings - International SoC Design Conference 2022, ISOCC 2022

SP - 129

EP - 130

BT - Proceedings - International SoC Design Conference 2022, ISOCC 2022

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 19th International System-on-Chip Design Conference, ISOCC 2022

Y2 - 19 October 2022 through 22 October 2022

ER -

The Quantitative Comparisons of Analog and Digital SRAM Compute-In-Memories for Deep Neural Network Applications

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this