Skip to content

Fix the inference of the function coloring for static compilation #340

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

Open
amontoison wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Fix the inference of the function coloring for static compilation #340

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2d2eafa
Update type_stability.jl
amontoison Sep 9, 2026
87db0e2
Fix the inference of the function coloring
amontoison Sep 6, 2026
d667043
Update test/type_stability.jl
amontoison Sep 10, 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
8 changes: 5 additions & 3 deletions src/interface.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -198,10 +198,12 @@ function coloring(
A::AbstractMatrix,
problem::ColoringProblem,
algo::GreedyColoringAlgorithm;
decompression_eltype::Type{R}=Float64,
decompression_eltype::Type=Float64,

@gdalle gdalle Sep 7, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The rationale behind the type annotation was to force specialization: https://docs.julialang.org/en/v1/manual/performance-tips/#Be-aware-of-when-Julia-avoids-specializing

Not sure how this change makes anything better in that regard, I would have thought it makes things worse?

@amontoison amontoison Sep 7, 2026 •
edited
Loading

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Me too and it is the contrary that is happening, the compiler widen the type to DataType internally from what I unserstand if we keep Type{R} and we don't have the inference in 3 cases.

I am wondering if it is because the argument with the specialization is a keyword argument.

symmetric_pattern::Bool=false,
) where {R}
return _coloring(WithResult(), A, problem, algo, R, symmetric_pattern)
)
return _coloring(
WithResult(), A, problem, algo, decompression_eltype, symmetric_pattern
)
end

"""
Expand Down
39 changes: 39 additions & 0 deletions test/type_stability.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,15 @@ using Test

rng = StableRNG(63)

# `@inferred` and `@test_opt` build their keyword arguments with `typeof`, which erases
# `Type{Float64}` down to `DataType`. Passing `decompression_eltype` straight to them would
# therefore hide the eltype from inference and make every result type unresolvable. Positional
# arguments go through `Core.Typeof` instead, which keeps `Type{R}` intact, so this wrapper is
# what lets us check that `coloring` stays inferrable when the eltype is given explicitly.
function coloring_with_eltype(A, problem, algo, ::Type{R}) where {R}
return coloring(A, problem, algo; decompression_eltype=R)
end

@testset "Sparse coloring" begin
n = 10
A = sparse(Symmetric(sprand(rng, n, n, 5 / n)))
Expand Down Expand Up @@ -57,6 +66,36 @@ rng = StableRNG(63)
)
end
end

@testset "Explicit decompression_eltype" begin
@testset "$structure - $partition - $decompression - $R" for (
structure, partition, decompression
) in [
(:nonsymmetric, :column, :direct),
(:nonsymmetric, :row, :direct),
(:symmetric, :column, :direct),
(:symmetric, :column, :substitution),
(:nonsymmetric, :bidirectional, :direct),
(:nonsymmetric, :bidirectional, :substitution),
],
R in (Float64, Float32)

@testset for order in all_orders()
@test_opt coloring_with_eltype(
A,
ColoringProblem(; structure, partition),
GreedyColoringAlgorithm(order; decompression),
R,
)
@inferred coloring_with_eltype(
A,
ColoringProblem(; structure, partition),
GreedyColoringAlgorithm(order; decompression),
R,
)
end
end
end
end;

@testset "Structured coloring" begin
Expand Down
Loading