MLIR: Machine Learning based IR Drop Prediction on ECO Revised Design for Faster Convergence

被引:11
作者
Kundu, Santanu [1 ]
Prasad, Manoranjan [1 ]
Nishad, Sashank [1 ]
Nachireddy, Sandeep [1 ]
Harikrishnan, K. [1 ]
机构
[1] Intel Technol India Pvt Ltd, SRR Campus, Bangalore, Karnataka, India
来源
2022 35TH INTERNATIONAL CONFERENCE ON VLSI DESIGN (VLSID 2022) HELD CONCURRENTLY WITH 2022 21ST INTERNATIONAL CONFERENCE ON EMBEDDED SYSTEMS (ES 2022) | 2022年
关键词
PDN convergence; Machine Learning; IR Drop Prediction; Explainable AI; Regression; Classification;
D O I
10.1109/VLSID2022.2022.00025
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The on-chip power delivery network (PDN) is an essential element of physical implementation that strongly determines functionality, quality, and reliability of an IC. During signoff phase, several engineering-change-order (ECO) iterations are needed to ensure that each instance of the design should meet IR drop specification. Even though the design remains very similar after ECO changes, conventional PDN analysis in industry standard CAD tool takes several hours of simulation runtime to determine IR drop. Our goal is to reduce this runtime in each iteration to evaluate the ECO changes and fix the violating cells immediately, prior to run conventional PDN signoff tool. Hence, improving the number of iteration and achieving faster PDN convergence. Our contribution in this paper is to develop a Machine Learning methodology for fast IR drop prediction where we have used regression techniques to predict static IR drop values and classification techniques to predict dynamic IR drop violating cells. We have evaluated the importance of every feature that contributes to the IR drop. We have also interpreted the predicted output using Explainable AI method. While inferencing on a similar to 1M instance industry design, we have achieved 0.997 R-2-Score, -5.21 mV maximum absolute error, and 117 mu V RMSE in static IR Drop prediction. On dynamic IR drop prediction, we have achieved 0.999 accuracy, 0.909 F1_Score, 0.893 Precision, and 0.926 Recall.
引用
收藏
页码:68 / 73
页数:6
相关论文
共 15 条
[1]  
[Anonymous], LIME
[2]  
[Anonymous], SCIKIT LEARN
[3]  
[Anonymous], IMBALANCE LEARN
[4]  
[Anonymous], FEATURE IMPORTANCES
[5]   SMOTE: Synthetic minority over-sampling technique [J].
Chawla, Nitesh V. ;
Bowyer, Kevin W. ;
Hall, Lawrence O. ;
Kegelmeyer, W. Philip .
2002, American Association for Artificial Intelligence (16)
[6]  
Chhabria V. A, 2021, PROC IEEE DATE
[7]   Thermal and IR Drop Analysis Using Convolutional Encoder-Decoder Networks [J].
Chhabria, Vidya A. ;
Ahuja, Vipul ;
Prabhu, Ashwath ;
Patil, Nikhil ;
Jain, Palkesh ;
Sapatnekar, Sachin S. .
2021 26TH ASIA AND SOUTH PACIFIC DESIGN AUTOMATION CONFERENCE (ASP-DAC), 2021, :690-696
[8]   Machine-learning-based Dynamic IR Drop Prediction for ECO [J].
Fang, Yen-Chun ;
Lin, Heng-Yi ;
Su, Min-Yan ;
Li, Chien-Mo ;
Fang, Eric Jia-Wei .
2018 IEEE/ACM INTERNATIONAL CONFERENCE ON COMPUTER-AIDED DESIGN (ICCAD) DIGEST OF TECHNICAL PAPERS, 2018,
[9]   IncPIRD: Fast Learning-Based Prediction of Incremental IR Drop [J].
Ho, Chia-Tung ;
Kahng, Andrew B. .
2019 IEEE/ACM INTERNATIONAL CONFERENCE ON COMPUTER-AIDED DESIGN (ICCAD), 2019,
[10]   Automatic IR-Drop ECO Using Machine Learning [J].
Lin, Heng-Yi ;
Fang, Yen-Chun ;
Liu, Shi-Tang ;
Chen, Jia-Xian ;
Li, Chien-Mo ;
Fang, Eric Jia-Wei .
2020 IEEE INTERNATIONAL TEST CONFERENCE IN ASIA (ITC-ASIA 2020), 2020, :7-12