Parallel FETI-DP algorithm for defect detection in large-area nanopatterned wafers

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

Abstract

An efficient parallelization of the dual-primal finite-element tearing and interconnecting algorithm (FETI-DP) is presented for detecting defects in large-area nanopatterned wafers. To reduce global communication, iterative solvers with an iterative classical Gram-Schmidt scheme and a communication-avoiding approach are implemented to solve the interface and the coarse corner problems, respectively. A preconditioner is developed to improve the iterative convergence of the coarse problem. Electromagnetic analysis of a wafer is presented to show the scalability and capability of the method for accurate defect detection.

Original languageEnglish (US)
Title of host publication2016 IEEE/ACES International Conference on Wireless Information Technology, ICWITS 2016 and System and Applied Computational Electromagnetics, ACES 2016 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509012596
DOIs
StatePublished - May 4 2016
EventIEEE/ACES International Conference on Wireless Information Technology, ICWITS 2016 and System and Applied Computational Electromagnetics, ACES 2016 - Honolulu, United States
Duration: Mar 13 2016Mar 17 2016

Publication series

Name2016 IEEE/ACES International Conference on Wireless Information Technology, ICWITS 2016 and System and Applied Computational Electromagnetics, ACES 2016 - Proceedings

Other

OtherIEEE/ACES International Conference on Wireless Information Technology, ICWITS 2016 and System and Applied Computational Electromagnetics, ACES 2016
Country/TerritoryUnited States
CityHonolulu
Period3/13/163/17/16

ASJC Scopus subject areas

  • Computational Mathematics
  • Signal Processing
  • Instrumentation
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Parallel FETI-DP algorithm for defect detection in large-area nanopatterned wafers'. Together they form a unique fingerprint.

Cite this