Efficient transpose of SparseMatrix

[markusrenoldner]

Dear members of the mfem community,

What is the standard way to explicitly compute the transpose of a
mfem::SparseMatrix
?

The class has an attribute *At, but it is proteced, i.e. it can not be accessed from a main() function. Further, the class offers a method called MultTranspose() which computes the matrix-vector-product with the transposed matrix. This does not help me, as I am assembling a larger matrix and therefore need explicit access to the transposed matrix.

Thank you for your help in advance.

[pazner]

You can use the function Transpose(const SparseMatrix&) which will return a new SparseMatrix* containing the transpose:

mfem/linalg/sparsemat.hpp

Lines 716 to 717 in b09cc38

	/// Transpose of a sparse matrix. A must be finalized.
	SparseMatrix *Transpose(const SparseMatrix &A);

[markusrenoldner]

Many thanks for the quick reply, this was very helpful!