Closed learning-chip closed 1 year ago
Reordering can reduce fill-in for both complete and incomplete factorizations. The CholeskyPreconditioner calls lldl() in LimitedLDLFactorizations.jl, but without suppling the reorder parameter:
CholeskyPreconditioner
lldl()
LimitedLDLFactorizations.jl
https://github.com/JuliaLinearAlgebra/Preconditioners.jl/blob/52b3702e9c8c5d42d6092ed2df7efdbe07c81411/src/incompletecholesky.jl#L8-L13
In lldl's example, AMD or METIS reordering can be used like:
AMD_P = amd(K1) METIS_P, METIS_invP = Metis.permutation(K) ... LLDL, t, b, g, m = @timed lldl(K1, Kdiag, AMD_P, memory = p) ... LLDL, t, b, g, m = @timed lldl(K1, Kdiag, METIS_P, memory = p)
A PR would be welcome :)
Fixed by #32.
Reordering can reduce fill-in for both complete and incomplete factorizations. The
CholeskyPreconditioner
callslldl()
inLimitedLDLFactorizations.jl
, but without suppling the reorder parameter:https://github.com/JuliaLinearAlgebra/Preconditioners.jl/blob/52b3702e9c8c5d42d6092ed2df7efdbe07c81411/src/incompletecholesky.jl#L8-L13
In lldl's example, AMD or METIS reordering can be used like: