Beyond low-order statistical response surfaces: Latent variable regression for efficient, highly nonlinear fitting

Amith Singhee, Rob A. Rutenbar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The number and magnitude of process variation sources are increasing as we scale further into the nano regime. Today's most successful response surface methods limit us to low-order forms - linear, quadratic - to make the fitting tractable. Unfortunately, not all variational scenarios are well modeled with low-order surfaces. We show how to exploit latent variable regression ideas to support efficient extraction of arbitrarily nonlinear statistical response surfaces. An implementation of these ideas called SiLVR, applied to a range of analog and digital circuits, in technologies from 90 to 45nm, shows significant improvements in prediction, with errors reduced by up to 21X, with very reasonable runtime costs.

Original languageEnglish (US)
Title of host publication2007 44th ACM/IEEE Design Automation Conference, DAC'07
Pages256-261
Number of pages6
DOIs
StatePublished - 2007
Event2007 44th ACM/IEEE Design Automation Conference, DAC'07 - San Diego, CA, United States
Duration: Jun 4 2007Jun 8 2007

Publication series

NameProceedings - Design Automation Conference
ISSN (Print)0738-100X

Other

Other2007 44th ACM/IEEE Design Automation Conference, DAC'07
CountryUnited States
CitySan Diego, CA
Period6/4/076/8/07

Keywords

  • DFM
  • Dimensionality reduction
  • Regression
  • Response surface

ASJC Scopus subject areas

  • Hardware and Architecture
  • Control and Systems Engineering

Fingerprint Dive into the research topics of 'Beyond low-order statistical response surfaces: Latent variable regression for efficient, highly nonlinear fitting'. Together they form a unique fingerprint.

Cite this