The Community for Technology Leaders
RSS Icon
Issue No.03 - May-June (2008 vol.25)
pp: 232-239
Pouria Bastani , University of California, Santa Barbara
Li-C. Wang , University of California, Santa Barbara
Magdy S. Abadir , Freescale Semiconductor
Traditional diagnosis of defects is based on an assumed fault model. A failing chip is diagnosed to find the subset of faults that can best explain the failure. This article discusses a new type of diagnosis to explain the mismatch between predicted timing behavior from modeling and simulation, and observed timing behavior measured on silicon. The authors illustrate that this type of diagnosis can be formulated as a statistical learning problem, and they propose a statistical diagnosis framework based on a learning technique called support vector classification. To diagnose the mismatch, they use a list of features to describe path characteristics. Each feature represents a potential source of uncertainty causing the mismatch. The output of the diagnosis is a rank of those features, such that a feature inducing a larger unexpected timing deviation is ranked higher. The authors explain the design of the proposed diagnosis framework, and they present experimental results to illustrate the effectiveness of the feature-ranking method.
statistical learning, timing mismatch, statistical diagnosis, feature ranking, timing behavior
Pouria Bastani, Li-C. Wang, Magdy S. Abadir, "Linking Statistical Learning to Diagnosis", IEEE Design & Test of Computers, vol.25, no. 3, pp. 232-239, May-June 2008, doi:10.1109/MDT.2008.79
1. M. Abramovici, M.A. Breuer, and A.D. Friedman, Digital Systems Testing and Testable Design, IEEE Press, 1990.
2. L.-C Wang, P. Bastani, and M.S. Abadir, "Design-Silicon Timing Correlation—A Data Mining Perspective," Proc. 44th Design Automation Conf. (DAC 07), ACM Press, 2007, pp. 384-389.
3. P. Bastani et al., "Analyzing the Risk of Timing Modeling Based on Path Delay Tests," Proc. Int'l Test Conf. (ITC 07), IEEE CS Press, 2007, paper 4437587 (10 pp.).
4. Z. Wang et al., "An Efficient and Effective Methodology on the Multiple Fault Diagnosis," Proc. Int'l Test Conf. (ITC 03), IEEE CS Press, 2003, pp. 329-338.
5. V.N. Vapnik, The Nature of Statistical Learning Theory, 2nd ed., Springer, 1999.
6. N. Cristianini and J. Shawe-Taylor, An Introduction to Support Vector Machines, Cambridge Univ. Press, 2002.
7. B. Schölkopf and A.J. Smola , , Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond, MIT Press, 2002.
18 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool