Bandwidth/fill-in minimzation algorithms #393
Labels
Comments
termi-official
changed the title
|
I have used Metis a bit: https://github.com/JuliaSparse/Metis.jl#fill-reducing-permutation |
fredrikekre
added a commit
that referenced
this issue
Dec 11, 2022
This patch adds new DoF renumbering possibilities using "package
extensions" which is available from Julia 1.10. These extensions hook
into the new `DofOrder.Ext{T}` type, where `T` is the external package
module providing the new ordering capability.
Currently only the fill-reducing renumbering algorithm from Metis.jl is
provided, ie.. `DofOrder.Ext{Metis}()`.
Closes #393.
fredrikekre
added a commit
that referenced
this issue
Dec 11, 2022
This patch adds new DoF renumbering possibilities using "package
extensions" which is available from Julia 1.10. These extensions hook
into the new `DofOrder.Ext{T}` type, where `T` is the external package
module providing the new ordering capability.
Currently only the fill-reducing renumbering algorithm from Metis.jl is
provided, ie.. `DofOrder.Ext{Metis}()`.
Closes #393.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Currently assembled matrices have a large bandwidth. This is quite problematic when utilizing direct solvers, because the memory consumption can grow out of hand for medium-sized problems. Showing how to utilize bandwidth minimization algorithms in some examples using direct solvers could be a good idea to populate best practices.
One package to start with here could be https://github.com/PetrKryslUCSD/SymRCM.jl which uses the reverse Cuthill-McKee reordering algorithm.
The text was updated successfully, but these errors were encountered: