The influence of machine learning on the future of static time analysis Julio Julio Torres Tello.
Article Sidebar
Main Article Content
Abstract
Introduction. Static Timing Analysis (STA), fundamental in the design of integrated circuits, involves evaluating the temporal performance of digital circuits under various conditions to meet certain constraints, through simulation. Despite its importance, traditional ATS faces several limitations when considering the increasing complexity of the integrated circuit manufacturing process in its models. The inclusion of Artificial Intelligence (AI) is seen as a promising solution to improve the precision and efficiency of the ATS, thus reducing design cycles in the electronics industry. Objective. To study the influence that the inclusion of AI has, and may have in the future, for the optimization of the ATS, and therefore to reduce design cycles in the electronics industry. Methodology. AI has been integrated into Static Time Analysis (STA), improving accuracy and efficiency when estimating delays, modeling process variations, and optimizing routes and synthesis processes. This integration addresses the complexity and variability of modern integrated circuits, accelerating design convergence, reducing iterations, and improving design quality. Additionally, AI is applied in model characterization for ATS, using adaptive simulations to accelerate the verification process and significantly reduce time to market, crucial in the semiconductor industry. This article reviews the current state and future projections of the contribution of AI in the ATS. Discussion. The future of STA promises a series of advances that seek to improve its capabilities and address emerging challenges in integrated circuit design. These developments include greater integration with Machine Learning (ML) to improve accuracy and efficiency. With the evolution towards smaller process nodes, STA will need to adapt to manage the increased complexity and variability introduced, employing more sophisticated ML algorithms. Additionally, STA is expected to focus more on power and reliability considerations, incorporating additional metrics and more complex data analysis, with the help of AI, to ensure energy efficiency and robustness against reliability issues. Conclusion. The future of ATS is shaping up to be constant innovation and adaptation to meet the changing needs of the semiconductor industry. Technological and methodological advances will play a crucial role in ensuring the timely delivery of high-performance and reliable integrated circuit designs. Given AI's data analysis and optimization capabilities, its revolutionary potential in ATS is considerable, especially with the growing inclusion of increasingly demanding requirements.
Downloads
Article Details
References
Castro-Zunti, R. D., Yepez, J., & Ko, S.-B. (2020). License plate segmentation and recognition system using deep learning and OpenVINO. IET Intelligent Transport Systems, 14(2): 119–126. https://doi.org/10.1049/iet-its.2019.0481
Chae, K. J., Jin, G. Y., Ko, S. B., Wang, Y., Zhang, H., Choi, E. J., & Choi, H. (2020). Deep learning for the classification of small (≤2 cm) pulmonary nodules on ct imaging: a preliminary study. Academic Radiology, 27(4), e55–e63. https://doi.org/10.1016/j.acra.2019.05.018
Chollet, F. (2017). Deep learning with python (1st edition). Manning Publications.
Design & Reuse. (s.f.). Integration of design-for-analysis in IC layout considerations to meet the challenges of shrinking technology. https://www.design-reuse.com/articles/18531/design-for-analysis-ic-layout.html
Forzan, C., & Pandini, D. (2009). Statistical static timing analysis: A survey. Integration, 42(3): 409–435. https://doi.org/10.1016/j.vlsi.2008.10.002
Géron, A. (2019). Hands-on machine learning with scikit-learn, Keras, and TensorFlow: concepts, tools, and techniques to build intelligent systems (2 edition). O’Reilly Media.
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. The MIT Press.
Han, Z., Zhao, J., Leung, H., Ma, K. F., & Wang, W. (2019). A review of deep learning models for time series prediction. IEEE Sensors Journal, 21(6): 7833-7848. https://doi.org/10.1109/JSEN.2019.2923982
Jiang, Z., Zhang, H., Wang, Y., & Ko, S.-B. (2018). Retinal blood vessel segmentation using fully convolutional network with transfer learning. Computerized Medical Imaging and Graphics, 68: 1–15. https://doi.org/10.1016/j.compmedimag.2018.04.005
Kaeslin, H. (Ed.). (2015). Chapter 4—Circuit modeling with hardware description languages. Top-Down Digital VLSI Design (pp. 179–300). Morgan Kaufmann. https://doi.org/10.1016/B978-0-12-800730-3.00004-6
Muthukrishnan, P., & Sathasivam, S. (2022). A technical survey on delay defects in nanoscale digital VLSI circuits. Applied Sciences, 12(18): 9103. https://doi.org/10.3390/app12189103
Raschka, S. (2015). Python Machine Learning, 1st Edition. Packt Publishing.
Tan, W.-L., & Santarini, M. (2023). How Siemens EDA’s Solid achieved production-grade AI in EDA applications. Siemens Digital Industries Software. https://resources.sw.siemens.com/en-US/white-paper-how-siemens-edas-solido-achieved-production-grade-ai-in-eda-applications
Torres-Tello, J., Guamán, A. V., & Ko, S.-B. (2020). Improving the detection of explosives in MOX chemical sensors array with LSTM Networks. IEEE Sensors Journal, 20(23): 14302–14309. https://doi.org/10.1109/JSEN.2020.3007431
Synopsys. (s.f.). What is Static Timing Analysis (STA)? – How STA works? https://www.synopsys.com/glossary/what-is-static-timing-analysis.html