make similar(A, [T,] shape...) for structured A yield a sparse rather than dense array #24170
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
base/linalg/lu.jl
Outdated
| @@ -416,6 +416,14 @@ function lufact!(A::Tridiagonal{T,V}, pivot::Union{Val{false}, Val{true}} = Val( | |||
| LU{T,Tridiagonal{T,V}}(B, ipiv, convert(BlasInt, info)) | |||
| end | |||
|
|
|||
| # Absent the following methods, if A::Tridiagonal{T} where !(T<:AbstractFloat), then the | |||
| # method lufact(A) hits sends copy!(similar(A, afloattype, size(A)), A) (which is not a | |||
| # Tridagonal given the (unnecessary?) shape argument to similar) to lufact! rather than | |||
There was a problem hiding this comment.
I think it is unnecessary so probably better to remove the size argument instead of introducing this method
There was a problem hiding this comment.
Nixing the shape argument sounds good. That removing the shape argument might break other consumers of the general lufact method motivated the present conservative solution. Let's see whether CI complains :). Best!
…er than dense array.
…ther than dense array.
|
For some reason, this triggers a test error in |
Should be fixed. Absent objections or requests for time and assuming CI clears, I plan to merge these changes this evening. (If someone beats me to it, not squashing the commits might be best for future bug sleuthing.) Best! |
Sacha0
force-pushed
the
similarstruc
branch
2 times, most recently
from
965b884
to
b3d12fd
Compare
…er than dense array.
Also fix a minor bug where lufact(A::Tridiagonal{T}) for !(T<:AbstractFloat) would dispatch to lufact!(B, ...) for non-tridiagonal rather than tridiagonal B downstream.
|
Thanks all! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This pull request makes
similar(A<:Union{Diagonal,Bidiagonal,Tridiagonal,SymTridiagonal}, [neweltype,] shape...)yield a sparse rather than dense array, consistent with e.g.broadcast. For example, this pull request makessimilar(Diagonal(rand(4)), Float32, (3, 3))yield aSparseMatrixCSC{Float32}rather than the presentMatrix{Float32}.This pull request also fixes a minor bug where
lufact(A<:Tridiagonal{T})where!(T<:AbstractFloat)would hitlufact!(B, ...)for non-Tridiagonalrather thanTridiagonalBdownstream.Ref. JuliaLang/LinearAlgebra.jl#271. Best!