Skip to content

Make GenericLUFactorization bypass overloads when used on DualLinear Problems #685

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
merged 8 commits into from
Aug 8, 2025
Merged

Make GenericLUFactorization bypass overloads when used on DualLinear Problems #685

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
fcc5eb4
add list of algs acceptable for going through Dual overloads
jClugstor Aug 7, 2025
b789400
make sure that GenericLUFactorization doesn't go through DualLinearCache
jClugstor Aug 7, 2025
ab0c6e4
add test
jClugstor Aug 7, 2025
532324a
fix test
jClugstor Aug 7, 2025
de4685d
make acceptable_algs in to const
jClugstor Aug 7, 2025
7f6b1e8
make opt out instead
jClugstor Aug 7, 2025
f05f2b9
make sure kwargs are passed in
jClugstor Aug 7, 2025
149029c
indirect import
jClugstor Aug 7, 2025
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
15 changes: 12 additions & 3 deletions ext/LinearSolveForwardDiffExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
module LinearSolveForwardDiffExt

using LinearSolve
using LinearSolve: SciMLLinearSolveAlgorithm
using LinearSolve: SciMLLinearSolveAlgorithm, __init
using LinearAlgebra
using ForwardDiff
using ForwardDiff: Dual, Partials
Expand Down Expand Up @@ -121,8 +121,17 @@ function linearsolve_dual_solution(u::AbstractArray, partials,
zip(u, partials_list[i, :] for i in 1:length(partials_list.u[1])))
end

function SciMLBase.init(
prob::DualAbstractLinearProblem, alg::LinearSolve.SciMLLinearSolveAlgorithm,
function SciMLBase.init(prob::DualAbstractLinearProblem, alg::SciMLLinearSolveAlgorithm, args...; kwargs...)
return __dual_init(prob, alg, args...; kwargs...)
end

# Opt out for GenericLUFactorization
function SciMLBase.init(prob::DualAbstractLinearProblem, alg::GenericLUFactorization, args...; kwargs...)
return __init(prob,alg, args...; kwargs...)
end

function __dual_init(
prob::DualAbstractLinearProblem, alg::SciMLLinearSolveAlgorithm,
args...;
alias = LinearAliasSpecifier(),
abstol = LinearSolve.default_tol(real(eltype(prob.b))),
Expand Down
6 changes: 5 additions & 1 deletion src/common.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -137,7 +137,11 @@ function __init_u0_from_Ab(A, b)
end
__init_u0_from_Ab(::SMatrix{S1, S2}, b) where {S1, S2} = zeros(SVector{S2, eltype(b)})

function SciMLBase.init(prob::LinearProblem, alg::SciMLLinearSolveAlgorithm,
function SciMLBase.init(prob::LinearProblem, alg::SciMLLinearSolveAlgorithm, args...; kwargs...)
__init(prob, alg, args...; kwargs...)
end

function __init(prob::LinearProblem, alg::SciMLLinearSolveAlgorithm,
args...;
alias = LinearAliasSpecifier(),
abstol = default_tol(real(eltype(prob.b))),
Expand Down
14 changes: 7 additions & 7 deletions test/forwarddiff_overloads.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,13 +87,6 @@ backslash_x_p = A \ new_b
@test ≈(x_p, backslash_x_p, rtol = 1e-9)

# Nested Duals
function h(p)
Copy link
Member

Choose a reason for hiding this comment

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

why was this removed?

Copy link
Member Author

Choose a reason for hiding this comment

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

It was a duplicate definition, it's first defined up above.

(A = [p[1] p[2]+1 p[2]^3;
3*p[1] p[1]+5 p[2] * p[1]-4;
p[2]^2 9*p[1] p[2]],
b = [p[1] + 1, p[2] * 2, p[1]^2])
end

A,
b = h([ForwardDiff.Dual(ForwardDiff.Dual(5.0, 1.0, 0.0), 1.0, 0.0),
ForwardDiff.Dual(ForwardDiff.Dual(5.0, 1.0, 0.0), 0.0, 1.0)])
Expand Down Expand Up @@ -193,3 +186,10 @@ overload_x_p = solve(prob, UMFPACKFactorization())
backslash_x_p = A \ b

@test ≈(overload_x_p, backslash_x_p, rtol = 1e-9)


# Test that GenericLU doesn't create a DualLinearCache
A, b = h([ForwardDiff.Dual(5.0, 1.0, 0.0), ForwardDiff.Dual(5.0, 0.0, 1.0)])

prob = LinearProblem(A, b)
@test init(prob, GenericLUFactorization()) isa LinearSolve.LinearCache
Loading