Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms

Hadi Esmaeilzadeh; Soroush Ghodrati; Andrew B. Kahng; Joon Kyung Kim; Sean Kinzer; Sayak Kundu; Rohan Mahapatra; Susmita Dey Manasi; Sachin S. Sapatnekar; Zhiang Wang; Ziqing Zeng

doi:10.1145/3551901.3556489

Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms

Hadi Esmaeilzadeh, Soroush Ghodrati, Andrew B. Kahng, Joon Kyung Kim, Sean Kinzer, Sayak Kundu, Rohan Mahapatra, Susmita Dey Manasi, Sachin S. Sapatnekar, Zhiang Wang, Ziqing Zeng

Electrical and Computer Engineering

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

2 Scopus citations

Abstract

Parameterizable ML accelerators are the product of recent breakthroughs in machine learning (ML). To fully enable the design space exploration, we propose a physical-design-driven, learning-based prediction framework for hardware-accelerated deep neural network (DNN) and non-DNN ML algorithms. It employs a unified methodology, coupling backend power, performance and area (PPA) analysis with frontend performance simulation, thus achieving realistic estimation of both backend PPA and system metrics (runtime and energy). Experimental studies show that the approach provides excellent predictions for both ASIC (in a 12nm commercial process) and FPGA implementations on the VTA and VeriGOOD-ML platforms.

Original language	English (US)
Title of host publication	MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD
Publisher	Association for Computing Machinery, Inc
Pages	119-126
Number of pages	8
ISBN (Electronic)	9781450394864
DOIs	https://doi.org/10.1145/3551901.3556489
State	Published - Sep 12 2022
Event	4th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2022 - Snowbird, United States Duration: Sep 12 2022 → Sep 13 2022

Publication series

Name	MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD

Conference

Conference	4th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2022
Country/Territory	United States
City	Snowbird
Period	9/12/22 → 9/13/22

Bibliographical note

Funding Information:
This material is based on research sponsored in part by Air Force Research Laboratory (AFRL) and Defense Advanced Research Projects Agency (DARPA) under agreement number FA8650-20-2-7009. Andrew B. Kahng is also supported in part by NSF CCF-2112665. The U. S. government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of AFRL, DARPA, or the U. S. government.

Publisher Copyright:
© 2022 Owner/Author.

Keywords

design space exploration
ML accelerator
PPA prediction

Access

10.1145/3551901.3556489

OpenUrl availability

Full text

Cite this

Esmaeilzadeh, H., Ghodrati, S., Kahng, A. B., Kim, J. K., Kinzer, S., Kundu, S., Mahapatra, R., Manasi, S. D., Sapatnekar, S. S., Wang, Z., & Zeng, Z. (2022). Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms. In MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD (pp. 119-126). (MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD). Association for Computing Machinery, Inc. https://doi.org/10.1145/3551901.3556489

Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms. / Esmaeilzadeh, Hadi; Ghodrati, Soroush; Kahng, Andrew B. et al.
MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD. Association for Computing Machinery, Inc, 2022. p. 119-126 (MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD).

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

Esmaeilzadeh, H, Ghodrati, S, Kahng, AB, Kim, JK, Kinzer, S, Kundu, S, Mahapatra, R, Manasi, SD, Sapatnekar, SS, Wang, Z & Zeng, Z 2022, Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms. in MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD. MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD, Association for Computing Machinery, Inc, pp. 119-126, 4th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2022, Snowbird, United States, 9/12/22. https://doi.org/10.1145/3551901.3556489

Esmaeilzadeh H, Ghodrati S, Kahng AB, Kim JK, Kinzer S, Kundu S et al. Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms. In MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD. Association for Computing Machinery, Inc. 2022. p. 119-126. (MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD). doi: 10.1145/3551901.3556489

Esmaeilzadeh, Hadi ; Ghodrati, Soroush ; Kahng, Andrew B. et al. / Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms. MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD. Association for Computing Machinery, Inc, 2022. pp. 119-126 (MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD).

@inproceedings{68064e63b1db4eacae4c826a93769021,

title = "Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms",

abstract = "Parameterizable ML accelerators are the product of recent breakthroughs in machine learning (ML). To fully enable the design space exploration, we propose a physical-design-driven, learning-based prediction framework for hardware-accelerated deep neural network (DNN) and non-DNN ML algorithms. It employs a unified methodology, coupling backend power, performance and area (PPA) analysis with frontend performance simulation, thus achieving realistic estimation of both backend PPA and system metrics (runtime and energy). Experimental studies show that the approach provides excellent predictions for both ASIC (in a 12nm commercial process) and FPGA implementations on the VTA and VeriGOOD-ML platforms.",

keywords = "design space exploration, ML accelerator, PPA prediction",

author = "Hadi Esmaeilzadeh and Soroush Ghodrati and Kahng, {Andrew B.} and Kim, {Joon Kyung} and Sean Kinzer and Sayak Kundu and Rohan Mahapatra and Manasi, {Susmita Dey} and Sapatnekar, {Sachin S.} and Zhiang Wang and Ziqing Zeng",

note = "Funding Information: This material is based on research sponsored in part by Air Force Research Laboratory (AFRL) and Defense Advanced Research Projects Agency (DARPA) under agreement number FA8650-20-2-7009. Andrew B. Kahng is also supported in part by NSF CCF-2112665. The U. S. government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of AFRL, DARPA, or the U. S. government. Publisher Copyright: {\textcopyright} 2022 Owner/Author.; 4th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2022 ; Conference date: 12-09-2022 Through 13-09-2022",

year = "2022",

month = sep,

day = "12",

doi = "10.1145/3551901.3556489",

language = "English (US)",

series = "MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD",

publisher = "Association for Computing Machinery, Inc",

pages = "119--126",

booktitle = "MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD",

}

TY - GEN

T1 - Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms

AU - Esmaeilzadeh, Hadi

AU - Ghodrati, Soroush

AU - Kahng, Andrew B.

AU - Kim, Joon Kyung

AU - Kinzer, Sean

AU - Kundu, Sayak

AU - Mahapatra, Rohan

AU - Manasi, Susmita Dey

AU - Sapatnekar, Sachin S.

AU - Wang, Zhiang

AU - Zeng, Ziqing

N1 - Funding Information: This material is based on research sponsored in part by Air Force Research Laboratory (AFRL) and Defense Advanced Research Projects Agency (DARPA) under agreement number FA8650-20-2-7009. Andrew B. Kahng is also supported in part by NSF CCF-2112665. The U. S. government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of AFRL, DARPA, or the U. S. government. Publisher Copyright: © 2022 Owner/Author.

PY - 2022/9/12

Y1 - 2022/9/12

N2 - Parameterizable ML accelerators are the product of recent breakthroughs in machine learning (ML). To fully enable the design space exploration, we propose a physical-design-driven, learning-based prediction framework for hardware-accelerated deep neural network (DNN) and non-DNN ML algorithms. It employs a unified methodology, coupling backend power, performance and area (PPA) analysis with frontend performance simulation, thus achieving realistic estimation of both backend PPA and system metrics (runtime and energy). Experimental studies show that the approach provides excellent predictions for both ASIC (in a 12nm commercial process) and FPGA implementations on the VTA and VeriGOOD-ML platforms.

AB - Parameterizable ML accelerators are the product of recent breakthroughs in machine learning (ML). To fully enable the design space exploration, we propose a physical-design-driven, learning-based prediction framework for hardware-accelerated deep neural network (DNN) and non-DNN ML algorithms. It employs a unified methodology, coupling backend power, performance and area (PPA) analysis with frontend performance simulation, thus achieving realistic estimation of both backend PPA and system metrics (runtime and energy). Experimental studies show that the approach provides excellent predictions for both ASIC (in a 12nm commercial process) and FPGA implementations on the VTA and VeriGOOD-ML platforms.

KW - design space exploration

KW - ML accelerator

KW - PPA prediction

UR - http://www.scopus.com/inward/record.url?scp=85139252276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85139252276&partnerID=8YFLogxK

U2 - 10.1145/3551901.3556489

DO - 10.1145/3551901.3556489

M3 - Conference contribution

AN - SCOPUS:85139252276

T3 - MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD

SP - 119

EP - 126

BT - MLCAD 2022 - Proceedings of the 2022 ACM/IEEE Workshop on Machine Learning for CAD

PB - Association for Computing Machinery, Inc

T2 - 4th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2022

Y2 - 12 September 2022 through 13 September 2022

ER -

Physically Accurate Learning-based Performance Prediction of Hardware-accelerated ML Algorithms

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this