Generating Routing-Driven Power Distribution Networks With Machine-Learning Technique

Wen Hsiang Chang; Chien Hsueh Lin; Szu Pang Mu; Li De Chen; Cheng Hong Tsai; Yen Chih Chiu; Chia-Tso Chao

doi:10.1109/TCAD.2017.2648842

Generating Routing-Driven Power Distribution Networks With Machine-Learning Technique

Wen Hsiang Chang^*, Chien Hsueh Lin, Szu Pang Mu, Li De Chen, Cheng Hong Tsai, Yen Chih Chiu, Chia-Tso Chao

^*Corresponding author for this work

Institute of Electronics

Research output: Contribution to journal › Article › peer-review

28 Scopus citations

Abstract

As technology node keeps scaling and design complexity keeps increasing, power distribution networks (PDNs) require more routing resource to meet IR-drop and electro-migration (EM) constraints. This paper presents a design flow to generate a PDN that can result in near-minimal overhead for the routing of the underlying standard cells while satisfying both IR-drop and EM constraints based on a given cell placement. The design flow relies on a machine-learning model to quickly predict the total wire length of global route associated with a given PDN configuration in order to speed up the search process. The experimental results based on various 28 nm industrial block designs have demonstrated the accuracy of the learned model for predicting the routing cost and the effectiveness of the proposed framework for reducing the routing cost of the final PDN.

Original language	English
Article number	7807300
Pages (from-to)	1237-1250
Number of pages	14
Journal	IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Volume	36
Issue number	8
DOIs	https://doi.org/10.1109/TCAD.2017.2648842
State	Published - Aug 2017

Keywords

Electro-migration (EM)
IR drop
machine learning
power grid design
routing cost model
routing-driven

Access to Document

10.1109/TCAD.2017.2648842

Cite this

@article{b8d1aea3e16546c9b2bab020b63c27c8,

title = "Generating Routing-Driven Power Distribution Networks With Machine-Learning Technique",

abstract = "As technology node keeps scaling and design complexity keeps increasing, power distribution networks (PDNs) require more routing resource to meet IR-drop and electro-migration (EM) constraints. This paper presents a design flow to generate a PDN that can result in near-minimal overhead for the routing of the underlying standard cells while satisfying both IR-drop and EM constraints based on a given cell placement. The design flow relies on a machine-learning model to quickly predict the total wire length of global route associated with a given PDN configuration in order to speed up the search process. The experimental results based on various 28 nm industrial block designs have demonstrated the accuracy of the learned model for predicting the routing cost and the effectiveness of the proposed framework for reducing the routing cost of the final PDN.",

keywords = "Electro-migration (EM), IR drop, machine learning, power grid design, routing cost model, routing-driven",

author = "Chang, {Wen Hsiang} and Lin, {Chien Hsueh} and Mu, {Szu Pang} and Chen, {Li De} and Tsai, {Cheng Hong} and Chiu, {Yen Chih} and Chia-Tso Chao",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.",

year = "2017",

month = aug,

doi = "10.1109/TCAD.2017.2648842",

language = "English",

volume = "36",

pages = "1237--1250",

journal = "IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems",

issn = "0278-0070",

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

number = "8",

}

TY - JOUR

T1 - Generating Routing-Driven Power Distribution Networks With Machine-Learning Technique

AU - Chang, Wen Hsiang

AU - Lin, Chien Hsueh

AU - Mu, Szu Pang

AU - Chen, Li De

AU - Tsai, Cheng Hong

AU - Chiu, Yen Chih

AU - Chao, Chia-Tso

PY - 2017/8

Y1 - 2017/8

N2 - As technology node keeps scaling and design complexity keeps increasing, power distribution networks (PDNs) require more routing resource to meet IR-drop and electro-migration (EM) constraints. This paper presents a design flow to generate a PDN that can result in near-minimal overhead for the routing of the underlying standard cells while satisfying both IR-drop and EM constraints based on a given cell placement. The design flow relies on a machine-learning model to quickly predict the total wire length of global route associated with a given PDN configuration in order to speed up the search process. The experimental results based on various 28 nm industrial block designs have demonstrated the accuracy of the learned model for predicting the routing cost and the effectiveness of the proposed framework for reducing the routing cost of the final PDN.

AB - As technology node keeps scaling and design complexity keeps increasing, power distribution networks (PDNs) require more routing resource to meet IR-drop and electro-migration (EM) constraints. This paper presents a design flow to generate a PDN that can result in near-minimal overhead for the routing of the underlying standard cells while satisfying both IR-drop and EM constraints based on a given cell placement. The design flow relies on a machine-learning model to quickly predict the total wire length of global route associated with a given PDN configuration in order to speed up the search process. The experimental results based on various 28 nm industrial block designs have demonstrated the accuracy of the learned model for predicting the routing cost and the effectiveness of the proposed framework for reducing the routing cost of the final PDN.

KW - Electro-migration (EM)

KW - IR drop

KW - machine learning

KW - power grid design

KW - routing cost model

KW - routing-driven

UR - http://www.scopus.com/inward/record.url?scp=85029362490&partnerID=8YFLogxK

U2 - 10.1109/TCAD.2017.2648842

DO - 10.1109/TCAD.2017.2648842

M3 - Article

AN - SCOPUS:85029362490

SN - 0278-0070

VL - 36

SP - 1237

EP - 1250

JO - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

JF - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

IS - 8

M1 - 7807300

ER -

Generating Routing-Driven Power Distribution Networks With Machine-Learning Technique

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this