Abstract

The method of arbitrary polynomial chaos (aPC) is used to characterize the delay and subthreshold leakage of commercial 65nm and 28nm standard cell libraries. Delay and subthreshold leakage of a standard cell are modelled as orthogonal polynomial series of process variables. The aPC approach constructs the polynomials using only moments, and the coefficients of the polynomials are determined by minimum least squares, using samples generated by Monte Carlo simulation (MCS). The mean and standard deviation of delay and subthreshold leakage computed using aPC are within 1% of those computed using MCS while the aPC method results in a 25X-40X speed up over MCS.

Original languageEnglish (US)
Title of host publicationLASCAS 2017 - 8th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference
Subtitle of host publicationProceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509058594
DOIs
StatePublished - Jun 13 2017
Event8th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2017 - Bariloche, Argentina
Duration: Feb 20 2017Feb 23 2017

Publication series

NameLASCAS 2017 - 8th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference: Proceedings

Other

Other8th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2017
CountryArgentina
CityBariloche
Period2/20/172/23/17

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Statistical library characterization using arbitrary polynomial chaos'. Together they form a unique fingerprint.

  • Cite this

    Ince, M., Ozev, S., & Vrudhula, S. (2017). Statistical library characterization using arbitrary polynomial chaos. In LASCAS 2017 - 8th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference: Proceedings [7948078] (LASCAS 2017 - 8th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference: Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/LASCAS.2017.7948078