2005
DOI: 10.1016/j.jcp.2004.10.021
|View full text |Cite
|
Sign up to set email alerts
|

Algebraic multigrid for higher-order finite elements

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
69
0

Year Published

2006
2006
2020
2020

Publication Types

Select...
8
1

Relationship

0
9

Authors

Journals

citations
Cited by 68 publications
(71 citation statements)
references
References 23 publications
2
69
0
Order By: Relevance
“…Also shown in the inset images are velocity vectors for specific branches. Because the triquartic finite element basis was based on Chebyshev-Lobatto node positions, which have been previously shown to work well with AMG solvers [20], the nodes are not evenly distributed but are somewhat clustered near the corners of elements. The global conservation of mass is approximately exact because C 3 was set to enforce this condition, but it is useful to examine the pressure gradient to determine the accuracy of the solution.…”
Section: Resultsmentioning
confidence: 99%
“…Also shown in the inset images are velocity vectors for specific branches. Because the triquartic finite element basis was based on Chebyshev-Lobatto node positions, which have been previously shown to work well with AMG solvers [20], the nodes are not evenly distributed but are somewhat clustered near the corners of elements. The global conservation of mass is approximately exact because C 3 was set to enforce this condition, but it is useful to examine the pressure gradient to determine the accuracy of the solution.…”
Section: Resultsmentioning
confidence: 99%
“…The use of the AMG methods for spectral elements has recently been studied in [34]. The number of connections between unknowns of the problem increases when higher-order elements are used.…”
Section: Preconditioned Conjugate Gradient Methodsmentioning
confidence: 99%
“…One way to obtain this matrix is to rediscretize the PDE with first-order finite elements while preserving the number and the location of the degrees of freedom [12,24]. Assuming that the element matrices are available as well as the corresponding assembly routine, another option is to compute a sparse approximation of each element matrix, and then assemble them to form the intermediate matrix [1].…”
Section: Introductionmentioning
confidence: 99%