Skip to content

Fix Jacobian sparsity pattern with observed derivatives #3949

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 2 commits into from
Oct 1, 2025
+18 −2
Merged

Fix Jacobian sparsity pattern with observed derivatives #3949

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7d56584
Test Jacobian sparsity with observed derivative
hersle Sep 30, 2025
16c9c0d
Disable computation of Jacobian sparsity pattern from torn system graph
hersle Sep 30, 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
5 changes: 3 additions & 2 deletions src/systems/codegen.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -455,8 +455,9 @@ end
Return the sparsity pattern of the jacobian of `sys` as a matrix.
"""
function jacobian_sparsity(sys::System)
sparsity = torn_system_jacobian_sparsity(sys)
sparsity === nothing || return sparsity
# disable to fix https://github.com/SciML/ModelingToolkit.jl/issues/3871
#sparsity = torn_system_jacobian_sparsity(sys)
#sparsity === nothing || return sparsity
Comment on lines -458 to +460
Copy link
Member

Choose a reason for hiding this comment

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

@AayushSabharwal is there a reason the torn sparsity would have false zeros here? That seems like it would give issues in other places.

Copy link
Member

Choose a reason for hiding this comment

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

I'm not 100% sure, we've just used torn_system_jacobian_sparsity since the beginning of time iirc. Maybe this doesn't handle the W sparsity correctly?

Copy link
Member

Choose a reason for hiding this comment

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

Well I would think that if this sparsity pattern is wrong, then the incidence matrix is wrong. Can you make a note to follow up on this? For now I want to merge so that these are right, but I'd like to make sure something isn't hiding here.

Copy link
Member

Choose a reason for hiding this comment

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

Will do


Symbolics.jacobian_sparsity([eq.rhs for eq in full_equations(sys)],
[dv for dv in unknowns(sys)])
Expand Down
15 changes: 15 additions & 0 deletions test/jacobiansparsity.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -139,3 +139,18 @@ end
@test J == prob.f.jac(prob.u0, prob.p, 1.0)
@test J ≈ prob.f.jac(prob.u0, prob.p, 1.0)
end

# https://github.com/SciML/ModelingToolkit.jl/issues/3871
@testset "Issue#3871: Sparsity with observed derivatives" begin
t = ModelingToolkit.t_nounits
D = ModelingToolkit.D_nounits
@variables x(t) y(t)
@mtkcompile sys = System([D(x) ~ x * D(y), D(y) ~ x - y], t)
@test ModelingToolkit.jacobian_sparsity(sys) == [1 1; 1 1] # all nonzero
J1 = calculate_jacobian(sys)
J2 = isequal(unknowns(sys)[1], x) ? [2x-y -x; 1 -1] : [-1 1; -x 2x-y] # analytical result
@test isequal(J1, J2)
prob = ODEProblem(sys, [x => 1.0, y => 0.0], (0.0, 1.0); jac = true, sparse = true)
sol = solve(prob, FBDF())
@test SciMLBase.successful_retcode(sol)
end
Loading