If it is not too expensive, I would copy the block vector to the normal
distributed vector. For this purpose, I use functions like these:
```
template <typename VectorType, typename BlockVectorType>
void
split_up_components_fast(const VectorType &src, BlockVectorType &dst)
{
for (unsigned int i = 0, j = 0;
i < src.get_partitioner()->locally_owned_size();
++j)
for (unsigned int b = 0; b < dst.n_blocks(); ++b)
dst.block(b).local_element(j) = src.local_element(i++);
}
template <typename VectorType, typename BlockVectorType>
void
merge_components_fast(const BlockVectorType &src, VectorType &dst)
{
for (unsigned int i = 0, j = 0;
i < dst.get_partitioner()->locally_owned_size();
++j)
for (unsigned int b = 0; b < src.n_blocks(); ++b)
dst.local_element(i++) = src.block(b).local_element(j);
}
```
However, you have make sure that the DoFs are enumerated consistently for
LinearAlgebra::distributed::Vector.
Peter
On Saturday, 5 November 2022 at 07:21:40 UTC+1 yy.wayne wrote:
> Hello,
>
> I assembled a matrix for a multi-component problem. I want to solve it
> with a direct solver, but neither Trilinos solver(SolverDirect amesos_klu)
> nor SparseDirectUMFPACK solver receive a
> LinearAlgebra::distributed::BlockVector as input. The Trilinos direct
> solver receives Vector and BlockVector, while SparseDirectUMFPACK solver
> receives LinearAlgebra::distributed::Vector. What direct solver should I
> use?
>
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to dealii+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/dealii/29454063-2eb4-4feb-9275-00a213ccac1an%40googlegroups.com.
