Statistical modeling of gate-delay variation with consideration of intra-gate variability

This paper proposes a model to calculate statistical gate-delay variation caused by intra-chip and inter-chip variabilities. The variation of each gate delay directly influences the variation of circuit delay, so it is important to characterize each gate-delay variation accurately. Our model characterizes the gate delay by transistor characteristics. Every transistor in a gate affects the transient characteristics of the gate, so it is indispensable to consider the intra-gate variability for the modeling of gate-delay variation. This effect is not captured in a statistical delay analysis reported so far. Our model characterizes a statistical gate-delay variation using a response surface method (RSM) and represents the intra-gate variability with a few parameters. We evaluate the accuracy of our model, and we show some simulated results of a circuit delay variation.

[1]  Sani R. Nassif,et al.  Modeling and analysis of manufacturing variations , 2001, Proceedings of the IEEE 2001 Custom Integrated Circuits Conference (Cat. No.01CH37169).

[2]  Hidetoshi Onodera,et al.  A statistical gate delay model for intra-chip and inter-chip variabilities , 2003, ASP-DAC '03.

[3]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[4]  George E. P. Box,et al.  Empirical Model‐Building and Response Surfaces , 1988 .

[5]  S.R. Nassif Within-chip variability analysis , 1998, International Electron Devices Meeting 1998. Technical Digest (Cat. No.98CH36217).

[6]  M.J.M. Pelgrom,et al.  Matching properties of MOS transistors , 1989 .

[7]  Costas J. Spanos,et al.  Circuit performance variability decomposition , 1999, 1999 4th International Workshop on Statistical Metrology (Cat. No.99TH8391).

[8]  E.T.A.F. Jacobs,et al.  Sources and quantification of delay variations in a 250nm CMOS digital cell library , 2000 .