Using subset_tunables leads to no method matching Float64(::ForwardDiff.Dual{...}) in some cases #3924
Description
Describe the bug 🐞
I have an example where using subset_tunables on a system leads to failures with ForwardDiff inside solve.
The interesting aspect is that due to the bug in #3921 we also have the original system with modified tunables, and if we use that one instead, we don't get errors and the results is the same as if we didn't use subset_tunables.
Expected behavior
No ForwardDiff issues.
Minimal Reproducible Example 👇
using ModelingToolkit
using ModelingToolkit: D_nounits as D, t_nounits as t
using SymbolicIndexingInterface
using OrdinaryDiffEqTsit5
using DifferentiationInterface: DifferentiationInterface as DI
using ADTypes, ForwardDiff
function create_sys()
@parameters p1 = 0.5 [tunable = true] (p23[1:2] = [1, 3.0]) [tunable = true] p4 = 3 * p1 [tunable = false] y0 = 1.2 [tunable = true]
@variables x(t) = 2p1 y(t) = y0 z(t) = x + y
eqs = [D(x) ~ p1 * x - p23[1] * x * y
D(y) ~ -p23[2] * y + p4 * x * y
z ~ x + y]
mtkcompile(System(eqs, t, name=:sys))
end
sys = create_sys()
sub_sys = subset_tunables(sys, [sys.p23])
tunable_parameters(sys)
prob = ODEProblem(sub_sys, [], (0, 1.)) # replace with sys
setter = setsym_oop(prob, Symbolics.scalarize(sys.p23));
function loss(x, ps)
setter, prob = ps
u0, p = setter(prob, x)
new_prob = remake(prob; u0, p)
sol = solve(new_prob, Tsit5())
sum(sol)
end
loss([1., 2], (setter, prob))
DI.gradient(Base.Fix2(loss, (setter, prob)), AutoForwardDiff(), [1, 2.])
# result without subset_tunables
# 2-element Vector{Float64}:
# -2.3921469699569458
# -1.1543421355772538Error & Stacktrace
ERROR: MethodError: no method matching Float64(::ForwardDiff.Dual{ForwardDiff.Tag{Base.Fix2{var"#loss", Tuple{…}}, Float64}, Float64, 2})
The type `Float64` exists, but no method is defined for this combination of argument types when trying to construct it.
Closest candidates are:
(::Type{T})(::Real, ::RoundingMode) where T<:AbstractFloat
@ Base rounding.jl:265
(::Type{T})(::T) where T<:Number
@ Core boot.jl:965
Float64(::Int16)
@ Base float.jl:245
...
Stacktrace:
[1] convert(::Type{Float64}, x::ForwardDiff.Dual{ForwardDiff.Tag{Base.Fix2{var"#loss", Tuple{…}}, Float64}, Float64, 2})
@ Base ./number.jl:7
[2] setindex!
@ ./array.jl:985 [inlined]
[3] macro expansion
@ ~/.julia/packages/SymbolicUtils/N76BL/src/code.jl:460 [inlined]
[4] macro expansion
@ ~/.julia/packages/Symbolics/FHhXE/src/build_function.jl:368 [inlined]
[5] macro expansion
@ ~/.julia/packages/RuntimeGeneratedFunctions/2SjTC/src/RuntimeGeneratedFunctions.jl:161 [inlined]
[6] macro expansion
@ ./none:0 [inlined]
[7] generated_callfunc
@ ./none:0 [inlined]
[8] (::RuntimeGeneratedFunctions.RuntimeGeneratedFunction{…})(::Vector{…}, ::Vector{…}, ::ModelingToolkit.MTKParameters{…}, ::Float64)
@ RuntimeGeneratedFunctions ~/.julia/packages/RuntimeGeneratedFunctions/2SjTC/src/RuntimeGeneratedFunctions.jl:148
[9] macro expansion
@ ~/.julia/packages/ModelingToolkit/H1QXB/src/systems/codegen_utils.jl:-1 [inlined]
[10] _generated_call(::ModelingToolkit.GeneratedFunctionWrapper{…}, ::Vector{…}, ::Vector{…}, ::ModelingToolkit.MTKParameters{…}, ::Float64)
@ ModelingToolkit ~/.julia/packages/ModelingToolkit/H1QXB/src/systems/codegen_utils.jl:353
[11] (::ModelingToolkit.GeneratedFunctionWrapper{…})(::Vector{…}, ::Vararg{…})
@ ModelingToolkit ~/.julia/packages/ModelingToolkit/H1QXB/src/systems/codegen_utils.jl:350
[12] (::SciMLBase.ODEFunction{…})(du::Vector{…}, u::Vector{…}, p::ModelingToolkit.MTKParameters{…}, t::Float64)
@ SciMLBase ~/.julia/packages/SciMLBase/pjTDe/src/scimlfunctions.jl:2592
[13] initialize!(integrator::OrdinaryDiffEqCore.ODEIntegrator{…}, cache::OrdinaryDiffEqTsit5.Tsit5Cache{…})
@ OrdinaryDiffEqTsit5 ~/.julia/packages/OrdinaryDiffEqTsit5/o07IY/src/tsit_perform_step.jl:175
[14] __init(prob::SciMLBase.ODEProblem{…}, alg::OrdinaryDiffEqTsit5.Tsit5{…}, timeseries_init::Tuple{}, ts_init::Tuple{}, ks_init::Tuple{}; saveat::Tuple{}, tstops::Tuple{}, d_discontinuities::Tuple{}, save_idxs::Nothing, save_everystep::Bool, save_on::Bool, save_discretes::Bool, save_start::Bool, save_end::Nothing, callback::Nothing, dense::Bool, calck::Bool, dt::Float64, dtmin::Float64, dtmax::Float64, force_dtmin::Bool, adaptive::Bool, gamma::Rational{…}, abstol::Nothing, reltol::Nothing, qmin::Rational{…}, qmax::Int64, qsteady_min::Int64, qsteady_max::Int64, beta1::Nothing, beta2::Nothing, qoldinit::Rational{…}, controller::Nothing, fullnormalize::Bool, failfactor::Int64, maxiters::Int64, internalnorm::typeof(DiffEqBase.ODE_DEFAULT_NORM), internalopnorm::typeof(LinearAlgebra.opnorm), isoutofdomain::typeof(DiffEqBase.ODE_DEFAULT_ISOUTOFDOMAIN), unstable_check::typeof(DiffEqBase.ODE_DEFAULT_UNSTABLE_CHECK), verbose::Bool, timeseries_errors::Bool, dense_errors::Bool, advance_to_tstop::Bool, stop_at_next_tstop::Bool, initialize_save::Bool, progress::Bool, progress_steps::Int64, progress_name::String, progress_message::typeof(DiffEqBase.ODE_DEFAULT_PROG_MESSAGE), progress_id::Symbol, userdata::Nothing, allow_extrapolation::Bool, initialize_integrator::Bool, alias::SciMLBase.ODEAliasSpecifier, initializealg::OrdinaryDiffEqCore.DefaultInit, kwargs::@Kwargs{})
@ OrdinaryDiffEqCore ~/.julia/packages/OrdinaryDiffEqCore/qzTLI/src/solve.jl:578
[15] __init (repeats 2 times)
@ ~/.julia/packages/OrdinaryDiffEqCore/qzTLI/src/solve.jl:11 [inlined]
[16] #__solve#51
@ ~/.julia/packages/OrdinaryDiffEqCore/qzTLI/src/solve.jl:6 [inlined]
[17] __solve
@ ~/.julia/packages/OrdinaryDiffEqCore/qzTLI/src/solve.jl:1 [inlined]
[18] #solve_call#21
@ ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:127 [inlined]
[19] solve_call(_prob::SciMLBase.ODEProblem{…}, args::OrdinaryDiffEqTsit5.Tsit5{…})
@ DiffEqBase ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:84
[20] solve_up(prob::SciMLBase.ODEProblem{…}, sensealg::Nothing, u0::Vector{…}, p::ModelingToolkit.MTKParameters{…}, args::OrdinaryDiffEqTsit5.Tsit5{…}; originator::SciMLBase.ChainRulesOriginator, kwargs::@Kwargs{})
@ DiffEqBase ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:683
[21] solve_up
@ ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:660 [inlined]
[22] solve(prob::SciMLBase.ODEProblem{…}, args::OrdinaryDiffEqTsit5.Tsit5{…}; sensealg::Nothing, u0::Nothing, p::Nothing, wrap::Val{…}, kwargs::@Kwargs{})
@ DiffEqBase ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:555
[23] solve(prob::SciMLBase.ODEProblem{…}, args::OrdinaryDiffEqTsit5.Tsit5{…})
@ DiffEqBase ~/.julia/packages/DiffEqBase/dqv41/src/solve.jl:545
[24] WARNING: Detected access to binding `Main.loss` in a world prior to its definition world.
Julia 1.12 has introduced more strict world age semantics for global bindings.
!!! This code may malfunction under Revise.
!!! This code will error in future versions of Julia.
Hint: Add an appropriate `invokelatest` around the access to this binding.
To make this warning an error, and hence obtain a stack trace, use `julia --depwarn=error`.
loss(x::Vector{ForwardDiff.Dual{…}}, ps::Tuple{SymbolicIndexingInterface.FullSetter{…}, SciMLBase.ODEProblem{…}})
@ Main ~/dev/subset_tunables_bug.jl:54
[25] Fix
@ ./operators.jl:1194 [inlined]
[26] vector_mode_dual_eval!
@ ~/.julia/packages/ForwardDiff/fKFOu/src/apiutils.jl:24 [inlined]
[27] vector_mode_gradient(f::Base.Fix2{…}, x::Vector{…}, cfg::ForwardDiff.GradientConfig{…})
@ ForwardDiff ~/.julia/packages/ForwardDiff/fKFOu/src/gradient.jl:98
[28] gradient
@ ~/.julia/packages/ForwardDiff/fKFOu/src/gradient.jl:20 [inlined]
[29] gradient(f::Base.Fix2{…}, x::Vector{…}, cfg::ForwardDiff.GradientConfig{…})
@ ForwardDiff ~/.julia/packages/ForwardDiff/fKFOu/src/gradient.jl:17
[30] gradient(f::Base.Fix2{var"#loss", Tuple{…}}, x::Vector{Float64})
@ ForwardDiff ~/.julia/packages/ForwardDiff/fKFOu/src/gradient.jl:17
[31] gradient(::Base.Fix2{var"#loss", Tuple{…}}, ::ADTypes.AutoForwardDiff{nothing, Nothing}, ::Vector{Float64})
@ DifferentiationInterfaceForwardDiffExt ~/.julia/packages/DifferentiationInterface/TuoGH/ext/DifferentiationInterfaceForwardDiffExt/onearg.jl:352
[32] top-level scopeEnvironment (please complete the following information):
- Output of
using Pkg; Pkg.status()
Status `/tmp/jl_nOYxND/Project.toml`
[47edcb42] ADTypes v1.17.0
[a0c0ee7d] DifferentiationInterface v0.7.7
[f6369f11] ForwardDiff v1.2.0
[961ee093] ModelingToolkit v10.22.0
[b1df2697] OrdinaryDiffEqTsit5 v1.5.0
[2efcf032] SymbolicIndexingInterface v0.3.43
- Output of
using Pkg; Pkg.status(; mode = PKGMODE_MANIFEST)
(jl_nOYxND) pkg> st -m
Status `/tmp/jl_nOYxND/Manifest.toml`
[47edcb42] ADTypes v1.17.0
[1520ce14] AbstractTrees v0.4.5
[7d9f7c33] Accessors v0.1.42
[79e6a3ab] Adapt v4.3.0
[66dad0bd] AliasTables v1.1.3
[ec485272] ArnoldiMethod v0.4.0
[4fba245c] ArrayInterface v7.20.0
[4c555306] ArrayLayouts v1.11.2
[e2ed5e7c] Bijections v0.2.2
[62783981] BitTwiddlingConvenienceFunctions v0.1.6
[8e7c35d0] BlockArrays v1.7.2
[70df07ce] BracketingNonlinearSolve v1.3.0
[2a0fbf3d] CPUSummary v0.2.7
[d360d2e6] ChainRulesCore v1.26.0
[fb6a15b2] CloseOpenIntervals v0.1.13
⌅ [861a8166] Combinatorics v1.0.2
[a80b9123] CommonMark v0.9.1
[38540f10] CommonSolve v0.2.4
[bbf7d656] CommonSubexpressions v0.3.1
[f70d9fcc] CommonWorldInvalidations v1.0.0
[34da2185] Compat v4.18.0
[b152e2b5] CompositeTypes v0.1.4
[a33af91c] CompositionsBase v0.1.2
[2569d6c7] ConcreteStructs v0.2.3
[187b0558] ConstructionBase v1.6.0
[adafc99b] CpuId v0.3.1
[9a962f9c] DataAPI v1.16.0
⌅ [864edb3b] DataStructures v0.18.22
[2b5f629d] DiffEqBase v6.189.1
[459566f4] DiffEqCallbacks v4.9.0
[77a26b50] DiffEqNoiseProcess v5.24.1
[163ba53b] DiffResults v1.1.0
[b552c78f] DiffRules v1.15.1
[a0c0ee7d] DifferentiationInterface v0.7.7
[8d63f2c5] DispatchDoctor v0.4.26
[31c24e10] Distributions v0.25.120
[ffbed154] DocStringExtensions v0.9.5
[5b8099bc] DomainSets v0.7.16
[7c1d4256] DynamicPolynomials v0.6.3
[06fc5a27] DynamicQuantities v1.8.0
[4e289a0a] EnumX v1.0.5
[f151be2c] EnzymeCore v0.8.13
[e2ba6199] ExprTools v0.1.10
[55351af7] ExproniconLite v0.10.14
[7034ab61] FastBroadcast v0.3.5
[9aa1b823] FastClosures v0.3.2
[a4df4552] FastPower v1.1.3
[1a297f60] FillArrays v1.14.0
[64ca27bc] FindFirstFunctions v1.4.2
[6a86dc24] FiniteDiff v2.28.1
[1fa38f19] Format v1.3.7
[f6369f11] ForwardDiff v1.2.0
[069b7b12] FunctionWrappers v1.1.3
[77dc65aa] FunctionWrappersWrappers v0.1.3
[d9f16b24] Functors v0.5.2
[46192b85] GPUArraysCore v0.2.0
[c27321d9] Glob v1.3.1
[86223c79] Graphs v1.13.1
[34004b35] HypergeometricFunctions v0.3.28
[615f187c] IfElse v0.1.1
[3263718b] ImplicitDiscreteSolve v1.2.0
[d25df0c9] Inflate v0.1.5
[18e54dd8] IntegerMathUtils v0.1.3
[8197267c] IntervalSets v0.7.11
[3587e190] InverseFunctions v0.1.17
[92d709cd] IrrationalConstants v0.2.4
[82899510] IteratorInterfaceExtensions v1.0.0
[692b3bcd] JLLWrappers v1.7.1
[ae98c720] Jieko v0.2.1
[98e50ef6] JuliaFormatter v2.1.6
⌅ [70703baa] JuliaSyntax v0.4.10
[ccbc3e58] JumpProcesses v9.19.1
[b964fa9f] LaTeXStrings v1.4.0
[23fbe1c1] Latexify v0.16.10
[10f19ff3] LayoutPointers v0.1.17
[87fe0de2] LineSearch v0.1.4
[d3d80556] LineSearches v7.4.0
[2ab3a3ac] LogExpFunctions v0.3.29
[d8e11817] MLStyle v0.4.17
[1914dd2f] MacroTools v0.5.16
[d125e4d3] ManualMemory v0.1.8
[bb5d69b7] MaybeInplace v0.1.4
[e1d29d7a] Missings v1.2.0
[961ee093] ModelingToolkit v10.22.0
[2e0e35c7] Moshi v0.3.7
[46d2c3a1] MuladdMacro v0.2.4
⌃ [102ac46a] MultivariatePolynomials v0.5.9
[d8a4904e] MutableArithmetics v1.6.4
[d41bc354] NLSolversBase v7.10.0
[77ba4419] NaNMath v1.1.3
[be0214bd] NonlinearSolveBase v1.14.0
[6fe1bfb0] OffsetArrays v1.17.0
[429524aa] Optim v1.13.2
[bac558e1] OrderedCollections v1.8.1
[bbf590c4] OrdinaryDiffEqCore v1.34.0
[b1df2697] OrdinaryDiffEqTsit5 v1.5.0
[90014a1f] PDMats v0.11.35
[d96e819e] Parameters v0.12.3
[e409e4f3] PoissonRandom v0.4.6
[f517fe37] Polyester v0.7.18
[1d0040c9] PolyesterWeave v0.2.2
[85a6dd25] PositiveFactorizations v0.2.4
[d236fae5] PreallocationTools v0.4.34
[aea7be01] PrecompileTools v1.3.3
[21216c6a] Preferences v1.5.0
[27ebfcd6] Primes v0.5.7
[43287f4e] PtrArrays v1.3.0
[1fd47b50] QuadGK v2.11.2
[74087812] Random123 v1.7.1
[e6cf234a] RandomNumbers v1.6.0
[3cdcf5f2] RecipesBase v1.3.4
[731186ca] RecursiveArrayTools v3.37.1
[189a3867] Reexport v1.2.2
[ae029012] Requires v1.3.1
[ae5879a3] ResettableStacks v1.1.1
[79098fc4] Rmath v0.8.0
[7e49a35a] RuntimeGeneratedFunctions v0.5.15
[9dfe8606] SCCNonlinearSolve v1.4.0
[94e857df] SIMDTypes v0.1.0
[0bca4576] SciMLBase v2.117.1
[19f34311] SciMLJacobianOperators v0.1.11
[c0aeaf25] SciMLOperators v1.7.1
[431bcebd] SciMLPublic v1.0.0
[53ae85a6] SciMLStructures v1.7.0
[efcf1570] Setfield v1.1.2
[727e6d20] SimpleNonlinearSolve v2.7.0
[699a6c99] SimpleTraits v0.9.5
[ce78b400] SimpleUnPack v1.1.0
[a2af1166] SortingAlgorithms v1.2.2
[276daf66] SpecialFunctions v2.5.1
[aedffcd0] Static v1.2.0
[0d7ed370] StaticArrayInterface v1.8.0
[90137ffa] StaticArrays v1.9.15
[1e83bf80] StaticArraysCore v1.4.3
[10745b16] Statistics v1.11.1
[82ae8749] StatsAPI v1.7.1
[2913bbd2] StatsBase v0.34.6
[4c63d2b9] StatsFuns v1.5.0
[7792a7ef] StrideArraysCore v0.5.8
[2efcf032] SymbolicIndexingInterface v0.3.43
[19f23fe9] SymbolicLimits v0.2.3
[d1185830] SymbolicUtils v3.32.0
[0c5d862f] Symbolics v6.54.0
[ed4db957] TaskLocalValues v0.1.3
[8ea1fca8] TermInterface v2.0.0
[1c621080] TestItems v1.0.0
[8290d209] ThreadingUtilities v0.5.5
[a759f4b9] TimerOutputs v0.5.29
[410a4b4d] Tricks v0.1.12
[781d530d] TruncatedStacktraces v1.4.0
[5c2747f8] URIs v1.6.1
[3a884ed6] UnPack v1.0.2
[1986cc42] Unitful v1.24.0
[a7c27f48] Unityper v0.1.6
[61579ee1] Ghostscript_jll v9.55.1+0
[aacddb02] JpegTurbo_jll v3.1.3+0
[efe28fd5] OpenSpecFun_jll v0.5.6+0
[f50d1b31] Rmath_jll v0.5.1+0
[56f22d72] Artifacts v1.11.0
[2a0f44e3] Base64 v1.11.0
[ade2ca70] Dates v1.11.0
[8ba89e20] Distributed v1.11.0
[9fa8497b] Future v1.11.0
[b77e0a4c] InteractiveUtils v1.11.0
[ac6e5ff7] JuliaSyntaxHighlighting v1.12.0
[8f399da3] Libdl v1.11.0
[37e2e46d] LinearAlgebra v1.12.0
[56ddb016] Logging v1.11.0
[d6f4376e] Markdown v1.11.0
[a63ad114] Mmap v1.11.0
[de0858da] Printf v1.11.0
[9a3f8284] Random v1.11.0
[ea8e919c] SHA v0.7.0
[9e88b42a] Serialization v1.11.0
[1a1011a3] SharedArrays v1.11.0
[6462fe0b] Sockets v1.11.0
[2f01184e] SparseArrays v1.12.0
[f489334b] StyledStrings v1.11.0
[4607b0f0] SuiteSparse
[fa267f1f] TOML v1.0.3
[8dfed614] Test v1.11.0
[cf7118a7] UUIDs v1.11.0
[4ec0a83e] Unicode v1.11.0
[e66e0078] CompilerSupportLibraries_jll v1.3.0+1
[4536629a] OpenBLAS_jll v0.3.29+0
[05823500] OpenLibm_jll v0.8.7+0
[bea87d4a] SuiteSparse_jll v7.8.3+2
[83775a58] Zlib_jll v1.3.1+2
[8e850b90] libblastrampoline_jll v5.13.1+1
Info Packages marked with ⌃ and ⌅ have new versions available. Those with ⌃ may be upgradable, but those with ⌅ are restricted by compatibility constraints from upgrading. To see why use `status --outdated -m`- Output of
versioninfo()
Julia Version 1.12.0-rc2
Commit 72cbf019d04 (2025-09-06 12:00 UTC)
Build Info:
Official https://julialang.org release
Platform Info:
OS: Linux (x86_64-linux-gnu)
CPU: 32 × Intel(R) Core(TM) i9-14900K
WORD_SIZE: 64
LLVM: libLLVM-18.1.7 (ORCJIT, alderlake)
GC: Built with stock GC
Threads: 32 default, 1 interactive, 32 GC (on 32 virtual cores)
Environment:
JULIA_EDITOR = code
JULIA_VSCODE_REPL = 1Additional context
I also tried this on #master, and the branches for #3923 & #3922.