Large-scale analog/RF performance modeling by statistical regression

Conference Paper

In this paper, we introduce several large-scale modeling techniques to analyze the high-dimensional, strongly-nonlinear performance variability observed in nanoscale manufacturing technologies. Our goal is to solve a large number of (e.g., 1044-106) model coefficients from a small set of (e.g., 102-103) sampling points without overfitting. This is facilitated by exploiting the underlying sparsity of model coefficients. Our circuit example designed in a commercial 65nm process demonstrates that the proposed techniques achieve 25x speedup compared with the traditional response surface modeling. ©2009 IEEE.

Full Text

Duke Authors

Cited Authors

  • Li, X

Published Date

  • December 1, 2009

Published In

  • Asicon 2009 Proceedings 2009 8th Ieee International Conference on Asic

Start / End Page

  • 646 - 649

International Standard Book Number 13 (ISBN-13)

  • 9781424438686

Digital Object Identifier (DOI)

  • 10.1109/ASICON.2009.5351329

Citation Source

  • Scopus