Skip to content

Try to do SVD truncation on GPU with _ind_intersect #148

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
lkdvos merged 6 commits into from
Jan 8, 2026
Merged

Try to do SVD truncation on GPU with _ind_intersect #148

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
c13ceca
Try to do SVD truncation on GPU with _ind_intersect
kshyatt Jan 8, 2026
fa9fa80
Use fixer GPUArrays branch
kshyatt Jan 8, 2026
aa8e07d
ind_intersect via filter
lkdvos Jan 8, 2026
a138765
Fix ambiguity
kshyatt Jan 8, 2026
076e943
Update src/implementations/truncation.jl
kshyatt Jan 8, 2026
e561986
Revert "Use fixer GPUArrays branch"
lkdvos Jan 8, 2026
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
7 changes: 3 additions & 4 deletions ext/MatrixAlgebraKitAMDGPUExt/MatrixAlgebraKitAMDGPUExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -167,9 +167,8 @@ function MatrixAlgebraKit._mul_herm!(C::StridedROCMatrix{T}, A::StridedROCMatrix
return C
end

# TODO: intersect on GPU arrays is not working
MatrixAlgebraKit._ind_intersect(A::ROCVector{Int}, B::AbstractVector) = MatrixAlgebraKit._ind_intersect(collect(A), B)
MatrixAlgebraKit._ind_intersect(A::AbstractVector, B::ROCVector{Int}) = MatrixAlgebraKit._ind_intersect(A, collect(B))
MatrixAlgebraKit._ind_intersect(A::ROCVector{Int}, B::ROCVector{Int}) = MatrixAlgebraKit._ind_intersect(collect(A), collect(B))
# TODO: intersect doesn't work on GPU
MatrixAlgebraKit._ind_intersect(A::ROCVector{Int}, B::ROCVector{Int}) =
MatrixAlgebraKit._ind_intersect(collect(A), collect(B))

end
7 changes: 3 additions & 4 deletions ext/MatrixAlgebraKitCUDAExt/MatrixAlgebraKitCUDAExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -191,9 +191,8 @@ function MatrixAlgebraKit._mul_herm!(C::StridedCuMatrix{T}, A::StridedCuMatrix{T
return C
end

# TODO: intersect on GPU arrays is not working
MatrixAlgebraKit._ind_intersect(A::CuVector{Int}, B::AbstractVector) = MatrixAlgebraKit._ind_intersect(collect(A), B)
MatrixAlgebraKit._ind_intersect(A::AbstractVector, B::CuVector{Int}) = MatrixAlgebraKit._ind_intersect(A, collect(B))
MatrixAlgebraKit._ind_intersect(A::CuVector{Int}, B::CuVector{Int}) = MatrixAlgebraKit._ind_intersect(collect(A), collect(B))
# TODO: intersect doesn't work on GPU
MatrixAlgebraKit._ind_intersect(A::CuVector{Int}, B::CuVector{Int}) =
MatrixAlgebraKit._ind_intersect(collect(A), collect(B))

end
9 changes: 9 additions & 0 deletions src/implementations/truncation.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -126,6 +126,15 @@ function _ind_intersect(A::AbstractVector{Bool}, B::AbstractVector)
end
_ind_intersect(A::AbstractVector, B::AbstractVector{Bool}) = _ind_intersect(B, A)
_ind_intersect(A::AbstractVector{Bool}, B::AbstractVector{Bool}) = A .& B

# when one of the ind selections is a unitrange, filter is more efficient than intersect
# since we know both selections only contain unique entries
# (This is also more GPU-friendly!)
_ind_intersect(A::AbstractUnitRange{Int}, B::AbstractUnitRange{Int}) = intersect(A, B)
_ind_intersect(A::AbstractVector{Int}, B::AbstractUnitRange{Int}) = filter(in(B), A)
_ind_intersect(A::AbstractUnitRange{Int}, B::AbstractVector{Int}) = _ind_intersect(B, A)

# when all else fails, call intersect
_ind_intersect(A, B) = intersect(A, B)

# Truncation error
Expand Down
Loading