-
-
Notifications
You must be signed in to change notification settings - Fork 5.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Dense * Tridiagonal = Sparse? #36551
Comments
I worked on these in a few older prs (I think #32521 is representative). I was told it was a conscious decision. It comes from how similar is defined for structured matrix types in sparsearrays. If this decision can be revisited I'd be quite happy to help. |
The issue is on which factor do you call julia> Mx*A
8×8 Array{Float64,2}:
-1.35375 -0.780296 0.512425 … -1.22988 -0.38358 -1.28576
1.27656 0.045418 -1.33725 0.503869 0.398921 0.772629
-1.67424 0.707068 0.813598 -0.389555 0.376913 -0.382437
0.948811 -0.33374 -0.0497121 -0.0619177 -1.10717 -0.612653
0.887939 -0.593686 0.176546 0.591326 0.928073 0.768298
-1.10373 0.670526 -0.522282 … -0.768006 -0.605545 0.569729
0.150751 -0.549392 0.0865425 0.468729 -0.0574477 -1.21706
-0.552833 0.0517309 -0.0678415 -0.239598 -0.411052 0.261905 gives dense output. If one could squeeze in better distinguishing multiplication methods without opening ambiguity hell, that would be great, but currently that's how it works. |
This seems to be a trap that will give you dense matrices in SparseMatrixCSC form. In general, Dense * Tridiagonal will be as dense as it started, so I'm not sure this was a conscious decision.
This was on v1.4, so I'm not sure if it's on master.
The text was updated successfully, but these errors were encountered: