Impact of element-level static condensation on iterative solver performance

D. Pardo, J. Álvarez-Aramberri, M. Paszynski, Lisandro Dalcin, Victor M. Calo

Research output: Contribution to journalArticlepeer-review

6 Scopus citations

Abstract

This paper provides theoretical estimates that quantify and clarify the savings associated to the use of element-level static condensation as a first step of an iterative solver. These estimates are verified numerically. The numerical evidence shows that static condensation at the element level is beneficial for higher-order methods. For lower-order methods or when the number of iterations required for convergence is low, the setup cost of the elimination as well as its implementation may offset the benefits obtained during the iteration process. However, as the iteration count (e.g., above 50) or the polynomial order (e.g., above cubics) grows, the benefits of element-level static condensation are significant.
Original languageEnglish (US)
Pages (from-to)2331-2341
Number of pages11
JournalComputers & Mathematics with Applications
Volume70
Issue number10
DOIs
StatePublished - Oct 5 2015

ASJC Scopus subject areas

  • Modeling and Simulation
  • Computational Theory and Mathematics
  • Computational Mathematics

Fingerprint

Dive into the research topics of 'Impact of element-level static condensation on iterative solver performance'. Together they form a unique fingerprint.

Cite this