Skip to content

can_setindex must be false for CUSPARSE arrays #495

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.

Sign up for GitHub

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

Merged
ChrisRackauckas merged 1 commit into from
Jul 14, 2026
+20 −0
Merged

can_setindex must be false for CUSPARSE arrays #495

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions ext/ArrayInterfaceCUDAExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ module ArrayInterfaceCUDAExt
using ArrayInterface
using CUDA
using CUDA.CUSOLVER
using CUDA.CUSPARSE
using LinearAlgebra

function ArrayInterface.lu_instance(A::CuMatrix{T}) where {T}
Expand All @@ -13,6 +14,17 @@ end

ArrayInterface.device(::Type{<:CUDA.CuArray}) = ArrayInterface.GPU()

# CUSPARSE arrays implement no `setindex!` at all, so the `true` default is wrong
# for them and callers guarding mutation with `can_setindex` hit a CanonicalIndexError.
const CuSparseArray = Union{
CUSPARSE.CuSparseVector,
CUSPARSE.CuSparseMatrixCSC,
CUSPARSE.CuSparseMatrixCSR,
CUSPARSE.CuSparseMatrixBSR,
CUSPARSE.CuSparseMatrixCOO,
}
ArrayInterface.can_setindex(::Type{<:CuSparseArray}) = false

function ArrayInterface.promote_eltype(
::Type{<:CUDA.CuArray{T, N, M}}, ::Type{T2}
) where {T, N, M, T2}
Expand Down
8 changes: 8 additions & 0 deletions test/gpu/cuda.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,3 +18,11 @@ lu_sparse = lu!(lu_inst_sparse, A_sparse)
#test that the resulting lu works
b = CuVector([1f0, 1f0])
@test CUDA.@allowscalar lu_sparse \ b == [1, 1]

# CUSPARSE arrays have no `setindex!`, so `can_setindex` must not claim they do
@test ArrayInterface.can_setindex(A_dense)
@test !ArrayInterface.can_setindex(A_sparse)
@test !ArrayInterface.can_setindex(CuSparseMatrixCSC(sparse(A_cpu)))
@test !ArrayInterface.can_setindex(CuSparseMatrixCOO(sparse(A_cpu)))
@test !ArrayInterface.can_setindex(CuSparseVector(sparsevec([1], [1.0f0], 2)))
@test_throws CanonicalIndexError A_sparse[1, 2] = 1.0f0
Loading