Low Cost Early Exit Decision Unit Design for CNN Accelerator

Geonho Kim; Jongsun Park

doi:10.1109/ISOCC50952.2020.9333079

Low Cost Early Exit Decision Unit Design for CNN Accelerator

Geonho Kim, Jongsun Park

School of Electrical Engineering

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

2 Citations (Scopus)

Abstract

Early exit has been studied as a way to reduce the complex computation of convolutional neural networks. However, in order to determine whether to exit early in a conventional CNN accelerator, there is a problem that a unit for computing softmax layer having a large hardware overhead is required. To solve this problem, we propose a low cost early exit decision unit. The proposed architecture uses only fully-connected (FC) layer outputs to make early exit decisions, so the computation of the softmax layer is not necessary. Our implementation results show an energy reduction of 68% with an accuracy drop of less than 0.3%.

Original language	English
Title of host publication	Proceedings - International SoC Design Conference, ISOCC 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	127-128
Number of pages	2
ISBN (Electronic)	9781728183312
DOIs	https://doi.org/10.1109/ISOCC50952.2020.9333079
Publication status	Published - 2020 Oct 21
Event	17th International System-on-Chip Design Conference, ISOCC 2020 - Yeosu, Korea, Republic of Duration: 2020 Oct 21 → 2020 Oct 24

Publication series

Name	Proceedings - International SoC Design Conference, ISOCC 2020

Conference

Conference	17th International System-on-Chip Design Conference, ISOCC 2020
Country/Territory	Korea, Republic of
City	Yeosu
Period	20/10/21 → 20/10/24

Keywords

CNN accelerator
early exit
softmax

ASJC Scopus subject areas

Energy Engineering and Power Technology
Electrical and Electronic Engineering
Instrumentation
Artificial Intelligence
Hardware and Architecture

Access to Document

10.1109/ISOCC50952.2020.9333079

Cite this

Low Cost Early Exit Decision Unit Design for CNN Accelerator. / Kim, Geonho; Park, Jongsun.
Proceedings - International SoC Design Conference, ISOCC 2020. Institute of Electrical and Electronics Engineers Inc., 2020. p. 127-128 9333079 (Proceedings - International SoC Design Conference, ISOCC 2020).

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

Kim, G & Park, J 2020, Low Cost Early Exit Decision Unit Design for CNN Accelerator. in Proceedings - International SoC Design Conference, ISOCC 2020., 9333079, Proceedings - International SoC Design Conference, ISOCC 2020, Institute of Electrical and Electronics Engineers Inc., pp. 127-128, 17th International System-on-Chip Design Conference, ISOCC 2020, Yeosu, Korea, Republic of, 20/10/21. https://doi.org/10.1109/ISOCC50952.2020.9333079

@inproceedings{af7805913fe3415982f001a5c8370cd1,

title = "Low Cost Early Exit Decision Unit Design for CNN Accelerator",

abstract = "Early exit has been studied as a way to reduce the complex computation of convolutional neural networks. However, in order to determine whether to exit early in a conventional CNN accelerator, there is a problem that a unit for computing softmax layer having a large hardware overhead is required. To solve this problem, we propose a low cost early exit decision unit. The proposed architecture uses only fully-connected (FC) layer outputs to make early exit decisions, so the computation of the softmax layer is not necessary. Our implementation results show an energy reduction of 68% with an accuracy drop of less than 0.3%. ",

keywords = "CNN accelerator, early exit, softmax",

author = "Geonho Kim and Jongsun Park",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 17th International System-on-Chip Design Conference, ISOCC 2020 ; Conference date: 21-10-2020 Through 24-10-2020",

year = "2020",

month = oct,

day = "21",

doi = "10.1109/ISOCC50952.2020.9333079",

language = "English",

series = "Proceedings - International SoC Design Conference, ISOCC 2020",

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

pages = "127--128",

booktitle = "Proceedings - International SoC Design Conference, ISOCC 2020",

}

TY - GEN

T1 - Low Cost Early Exit Decision Unit Design for CNN Accelerator

AU - Kim, Geonho

AU - Park, Jongsun

PY - 2020/10/21

Y1 - 2020/10/21

N2 - Early exit has been studied as a way to reduce the complex computation of convolutional neural networks. However, in order to determine whether to exit early in a conventional CNN accelerator, there is a problem that a unit for computing softmax layer having a large hardware overhead is required. To solve this problem, we propose a low cost early exit decision unit. The proposed architecture uses only fully-connected (FC) layer outputs to make early exit decisions, so the computation of the softmax layer is not necessary. Our implementation results show an energy reduction of 68% with an accuracy drop of less than 0.3%.

AB - Early exit has been studied as a way to reduce the complex computation of convolutional neural networks. However, in order to determine whether to exit early in a conventional CNN accelerator, there is a problem that a unit for computing softmax layer having a large hardware overhead is required. To solve this problem, we propose a low cost early exit decision unit. The proposed architecture uses only fully-connected (FC) layer outputs to make early exit decisions, so the computation of the softmax layer is not necessary. Our implementation results show an energy reduction of 68% with an accuracy drop of less than 0.3%.

KW - CNN accelerator

KW - early exit

KW - softmax

UR - http://www.scopus.com/inward/record.url?scp=85100773253&partnerID=8YFLogxK

U2 - 10.1109/ISOCC50952.2020.9333079

DO - 10.1109/ISOCC50952.2020.9333079

M3 - Conference contribution

AN - SCOPUS:85100773253

T3 - Proceedings - International SoC Design Conference, ISOCC 2020

SP - 127

EP - 128

BT - Proceedings - International SoC Design Conference, ISOCC 2020

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 17th International System-on-Chip Design Conference, ISOCC 2020

Y2 - 21 October 2020 through 24 October 2020

ER -

Low Cost Early Exit Decision Unit Design for CNN Accelerator

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this