A special case for the caching ExpRK methods is when the linear operator is diagonal. For example, this occurs when you have Fourier decompositions of a PDE operator with periodic BCs. The current implementation of the integrators treat all matrices in a generic way, and it would help to have special treatment for the diagonal case.
MSeeker1340
commented
6 years ago
A special case for the caching ExpRK methods is when the linear operator is diagonal. For example, this occurs when you have Fourier decompositions of a PDE operator with periodic BCs. The current implementation of the integrators treat all matrices in a generic way, and it would help to have special treatment for the diagonal case.