MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs

Yishuang Lin; Yaguang Li; Meghna Madhusudan; Sachin S. Sapatnekar; Ramesh Harjani; Jiang Hu

doi:10.1109/MLCAD58807.2023.10299829

MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs

Yishuang Lin
, Yaguang Li
, Meghna Madhusudan
, Sachin S. Sapatnekar
, Ramesh Harjani
, Jiang Hu

Electrical and Computer Engineering

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

1 Scopus citations

Abstract

Performance modeling is a key bottleneck for analog design automation. Although machine learning-based models have advanced the state-of-the-art, they have so far suffered from huge data preparation cost, very limited reusability, and inadequate accuracy for large circuits. We introduce ML-based macro-modeling techniques to mitigate these problems for linear analog ICs and ADC/DACs. On representative testcases, our method achieves more than 1700× speedup for data preparation and remarkably smaller model errors compared to recent ML approaches. It also attains 3600× acceleration over SPICE simulation with very small errors and reduces data preparation time for an ADC design from 40 days to 9.6 hours.

Original language	English (US)
Title of host publication	2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350309553
DOIs	https://doi.org/10.1109/MLCAD58807.2023.10299829
State	Published - 2023
Event	5th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2023 - Snowbird, United States Duration: Sep 10 2023 → Sep 13 2023

Publication series

Name	2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023

Conference

Conference	5th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2023
Country/Territory	United States
City	Snowbird
Period	9/10/23 → 9/13/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Publisher link

10.1109/MLCAD58807.2023.10299829

Find It

Find It at U of M Libraries

Cite this

Lin, Y., Li, Y., Madhusudan, M., Sapatnekar, S. S., Harjani, R., & Hu, J. (2023). MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs. In 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023 (2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/MLCAD58807.2023.10299829

MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs. / Lin, Yishuang; Li, Yaguang; Madhusudan, Meghna et al.
2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023).

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

Lin, Y, Li, Y, Madhusudan, M, Sapatnekar, SS , Harjani, R & Hu, J 2023, MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs. in 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023. 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023, Institute of Electrical and Electronics Engineers Inc., 5th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2023, Snowbird, United States, 9/10/23. https://doi.org/10.1109/MLCAD58807.2023.10299829

Lin Y, Li Y, Madhusudan M, Sapatnekar SS , Harjani R, Hu J. MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs. In 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023. Institute of Electrical and Electronics Engineers Inc. 2023. (2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023). doi: 10.1109/MLCAD58807.2023.10299829

@inproceedings{6eb07d0191644c65997289772f2fe442,

title = "MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs",

abstract = "Performance modeling is a key bottleneck for analog design automation. Although machine learning-based models have advanced the state-of-the-art, they have so far suffered from huge data preparation cost, very limited reusability, and inadequate accuracy for large circuits. We introduce ML-based macro-modeling techniques to mitigate these problems for linear analog ICs and ADC/DACs. On representative testcases, our method achieves more than 1700{\texttimes} speedup for data preparation and remarkably smaller model errors compared to recent ML approaches. It also attains 3600{\texttimes} acceleration over SPICE simulation with very small errors and reduces data preparation time for an ADC design from 40 days to 9.6 hours.",

author = "Yishuang Lin and Yaguang Li and Meghna Madhusudan and Sapatnekar, \{Sachin S.\} and Ramesh Harjani and Jiang Hu",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 5th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2023 ; Conference date: 10-09-2023 Through 13-09-2023",

year = "2023",

doi = "10.1109/MLCAD58807.2023.10299829",

language = "English (US)",

series = "2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023",

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

booktitle = "2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023",

}

TY - GEN

T1 - MMM

T2 - 5th ACM/IEEE Workshop on Machine Learning for CAD, MLCAD 2023

AU - Lin, Yishuang

AU - Li, Yaguang

AU - Madhusudan, Meghna

AU - Sapatnekar, Sachin S.

AU - Harjani, Ramesh

AU - Hu, Jiang

PY - 2023

Y1 - 2023

N2 - Performance modeling is a key bottleneck for analog design automation. Although machine learning-based models have advanced the state-of-the-art, they have so far suffered from huge data preparation cost, very limited reusability, and inadequate accuracy for large circuits. We introduce ML-based macro-modeling techniques to mitigate these problems for linear analog ICs and ADC/DACs. On representative testcases, our method achieves more than 1700× speedup for data preparation and remarkably smaller model errors compared to recent ML approaches. It also attains 3600× acceleration over SPICE simulation with very small errors and reduces data preparation time for an ADC design from 40 days to 9.6 hours.

AB - Performance modeling is a key bottleneck for analog design automation. Although machine learning-based models have advanced the state-of-the-art, they have so far suffered from huge data preparation cost, very limited reusability, and inadequate accuracy for large circuits. We introduce ML-based macro-modeling techniques to mitigate these problems for linear analog ICs and ADC/DACs. On representative testcases, our method achieves more than 1700× speedup for data preparation and remarkably smaller model errors compared to recent ML approaches. It also attains 3600× acceleration over SPICE simulation with very small errors and reduces data preparation time for an ADC design from 40 days to 9.6 hours.

UR - https://www.scopus.com/pages/publications/85178551858

UR - https://www.scopus.com/pages/publications/85178551858#tab=citedBy

U2 - 10.1109/MLCAD58807.2023.10299829

DO - 10.1109/MLCAD58807.2023.10299829

M3 - Conference contribution

AN - SCOPUS:85178551858

T3 - 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023

BT - 2023 ACM/IEEE 5th Workshop on Machine Learning for CAD, MLCAD 2023

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 10 September 2023 through 13 September 2023

ER -

MMM: Machine Learning-Based Macro-Modeling for Linear Analog ICs and ADC/DACs

Abstract

Publication series

Conference

Bibliographical note

Publisher link

Find It

Other files and links

Fingerprint

Cite this