Start adding support for matrix functions #135
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Start adding support for matrix functions such as
exp(A)
,cos(A)
,inv(A)
, etc. on square block diagonal matrices. The operations are applied to the diagonal blocks. Care is taken to instantiate unstored diagonal blocks in case they are nontrivial (or even if they remain zero they get instantiated for simplicity).Matrix functions supported by Julia can be found:
The strategy to handle this generically is to use
Base.promote_op
to use type inference on the block type of the input block sparse matrix to determine the block type of the output.A number of operations are broken right now because the LinearAlgebra.jl dense implementations are type unstable, I'll look into fixing those manually. Maybe we can provide an interface likeThis issue is fixed by introducinginitialize_output(f, A)
(inspired byMatrixAlgebraKit.initialize_output
) for customizing/specifying the output type. It can default toBase.promote_op
.initialize_output_blocksparse(f, A)
, which defaults toBase.promote_op
but then usessimilartype
in cases that are type unstable.