臺灣博碩士論文加值系統

在先進的IC 設計領域，過多的電路壓降會降低電路性能、導致時序違規甚至導致功能故障。因此，電路壓降的問題一直是電源供應網路分析中的主要問題，並受到了相當多的關注。本論文的目的是開發一個快速且準確的電路壓降預測器，以減少節點分析的設計時間並直接獲得節點不符合電路壓降的分析結果。在此引用支持向量機，並從電源供應網路模型中萃取具有影響力的特徵來訓練機器學習之模型。在構建機器學習模型時，我們使用一些技巧來避免過擬合的狀況發生。我們將此研究應用在台積電180 奈米製程下的工業設計，實驗結果表示，此模型能夠以98.11％至9.42％的高準確度有效預測電源壓降是否符合目標限制，並同時討論特徵的影響及重要性。

In the advanced IC design, excessive IR drop will slow down the circuit performance, causing timing violation and even leading to function failure. Therefore, IR drop problem has been a major issue in the analysis of power distribution network and has received great attention.
The objective of this work is to develop a fast and accurate IR drop predictor to reduce runtime of nodal analysis and obtain the IR drop violation result directly from the power delivery network (PDN). Support vector machine (SVM) is applied and it helps train influential features extracted from PDN model to predict IR drop result. When constructing the machine learning model, we use some skills to avoid overfitting and tune the parameter to find better performance. Our work is experimented on a real industry design in TSMC 180 nm process. Experimental result shows our model can efficiently predict IR violation with high accuracy of 98.11% to 99.42% and at the same time discuss the influential features.

Abstract (Traditional Chinese) . . . . . . . . . . . . . . . . . . . . . . . . . i
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Previous Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Organization of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Chapter 2 Preliminary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Prototype of PDN Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Determine Location of Stripe by Clustering Method . . . . . . . . . . . . . 6
2.3 Support Vector Machines (SVM) . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Kernel Function of SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Chapter 3 Overview of PDN Generation and Design Flow . . . . . . . . 14
Chapter 4 Machine-Learning Based Modeling . . . . . . . . . . . . . . . 16
4.1 Features Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Training Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2.1 Overfitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2.2 Training Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2.3 Training Skill to Improve Performance . . . . . . . . . . . . . . . . 22
4.3 Classification Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Chapter 5 Experimental Result . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.2 Experiment 1: Model Accuracy . . . . . . . . . . . . . . . . . . . . . . . . 26
5.3 Experiment 2: Different Distribution of PDN . . . . . . . . . . . . . . . . 30
5.4 Experiment 3: Different Kernel Function . . . . . . . . . . . . . . . . . . . 31
5.5 Experiment 4: Influence of Feature . . . . . . . . . . . . . . . . . . . . . . 32
v
Chapter 6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

[1] Y. Chuang, C. Lin, H. Chen, C. Lee, and T. Chen, “More effective power network prototyping by analytical and centroid learning,” in 2019 IEEE International Symposium on Circuits and Systems (ISCAS), pp. 1–5, 2019.
[2] W. Liao, C. Lin, S. Fang, C. Huang, H. Chen, D. Kwai, and Y. Chou, “Heterogeneous chip power delivery modeling and co-synthesis for practical 3DIC realization,” in 2017 22nd Asia and South Pacific Design Automation Conference (ASP-DAC), pp. 549–553, 2017.
[3] S. Köse and E. G. Friedman, “Fast algorithms for IR voltage drop analysis exploiting locality,” in 2011 48th ACM/EDAC/IEEE Design Automation Conference (DAC), pp. 996–1001, 2011.
[4] Z. Zeng, Z. Feng, P. Li, and V. Sarin, “Locality-driven parallel static analysis for power delivery networks,” in ACM Transactions on Design Automation of Electronic Systems, no. 28, 2011.
[5] E. Chiprout, “Fast flip-chip power grid analysis via locality and grid shells,” in IEEE/ACM International Conference on Computer Aided Design, 2004. ICCAD-2004, pp. 485–488, 2004.
[6] C. Pao, A. Su, and Y. Lee, “XGBIR: an XGBoost-based IR drop predictor for power delivery network,” in 2020 Design, Automation Test in Europe Conference Exhibition (DATE), pp. 1307–1310, 2020.
[7] C. Ho and A. B. Kahng, “IncPIRD: fast learning-based prediction of incremental IR drop,” in 2019 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), pp. 1–8, 2019.
[8] Y. Fang, H. Lin, M. Sui, C. Li, and E. J. Fang, “Machine-learning-based dynamic IR drop prediction for ECO,” in 2018 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), pp. 1–7, 2018.
[9] S. Lin, Y. Fang, Y. Li, Y. Liu, T. Yang, S. Lin, C. Li, and E. J. Fang, “IR drop prediction of ECO-revised circuits using machine learning,” in 2018 IEEE 36th VLSI Test Symposium (VTS), pp. 1–6, 2018.
[10] Z. Xie, H. Ren, B. Khailany, Y. Sheng, S. Santosh, J. Hu, and Y. Chen, “PowerNet: transferable dynamic IR drop estimation via maximum convolutional neural network,” in 2020 25th Asia and South Pacific Design Automation Conference (ASPDAC), pp. 13–18, 2020.
[11] L. Chen, C. Huang, Y. Chang, and H. Chen, “A learning-based methodology for routability prediction in placement,” in 2018 International Symposium on VLSI Design, Automation and Test (VLSI-DAT), pp. 1–4, 2018.
[12] S. S. Liu, C. Lee, C. Huang, H. Chen, C. Lin, and C. Lee, “Effective power network prototyping via statistical-based clustering and sequential linear programming,” in 2013 Design, Automation Test in Europe Conference Exhibition (DATE), pp. 1701–1706, 2013.
[13] C. Huang, C. Lin, W. Liao, C. Lee, H. Chen, C. Lee, and D. Kwai, “Improving power delivery network design by practical methodologies,” in 2014 IEEE 32nd International Conference on Computer Design (ICCD), pp. 237–242, 2014.
[14] S. Lloyd, “Least squares quantization in PCM,” IEEE Transactions on Information Theory, vol. 28, no. 2, pp. 129–137, 1982.
[15] C.-C. Chang and C.-J. Lin, “LIBSVM: A library for support vector machines,” ACM Trans. Intell. Syst. Technol., vol. 2, no. 3, p. 27:1–27:27, 2011.
[16] M. Guiney and E. Leavitt, “An introduction to OpenAccess an open source data model and API for IC design,” in Asia and South Pacific Conference on Design Automation, 2006, pp. 434–436, 2006.
[17] G. Venezian, “On the resistance between two points on a grid,” American Journal of Physics, vol. 62, pp. 1000–1004, 1994.
[18] S. Kose and E. G. Friedman, “Effective resistance of a two layer mesh,” IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 58, no. 11, pp. 739–743, 2011.
[19] H. He and Y. Ma, Imbalanced Learning: Foundations, Algorithms, and Applications, 1st ed. Wiley-IEEE Press, 2013.
[20] IBM ILOG CPLEX,
https://www.ibm.com/products/software.”