Mutating operations in ODE-like problem encounters Illegal replace ficticious phi error for Julia 1.11 but not Julia 1.10.1

[xkykai]

Hello, I have encountered an issue when trying to perform AD on an ODE-like problem. Below is my MWE:

using Enzyme

ps = rand(32)

function solve_equation(ps, u0)
    Δt = 1e-1
    sol = zeros(32, 3)
    sol[:, 1] .= u0
    u = u0
    for i in 2:3
        sol[:, i] .= u .+ Δt .* ps
        u .= sol[:, i]
    end

    return sol
end

truth = rand(32, 3)
u0 = truth[:, 1]

solve_equation(ps, u0)

function loss(ps, truth, u0)
    return sum(abs2, solve_equation(ps, u0) .- truth)
end

loss(ps, truth, u0)

dps = deepcopy(ps) .= 0
autodiff(Enzyme.Reverse, loss, Active, Duplicated(ps, dps), Const(truth), Duplicated(u0, deepcopy(u0)))

This gives an error of

ERROR: Enzyme compilation failed due to an internal error.
 Please open an issue with the code to reproduce and full error log on github.com/EnzymeAD/Enzyme.jl
 To toggle more information for debugging (needed for bug reports), set Enzyme.Compiler.VERBOSE_ERRORS[] = true (default false)

Illegal replace ficticious phi for:   %_replacementA732 = phi i8* , !dbg !347 of   %335 = load i8*, i8* addrspace(11)* %334, align 8, !dbg !314, !tbaa !33, !alias.scope !186, !noalias !187, !nonnull !0, !enzyme_type !61, !enzymejl_byref_BITS_VALUE !0, !enzymejl_source_type_Ptr\7BFloat64\7D !0

Stacktrace:
  [1] memoryref
    @ .\boot.jl:522
  [2] copy
    @ .\array.jl:351
  [3] unaliascopy
    @ .\abstractarray.jl:1516
  [4] unalias
    @ .\abstractarray.jl:1500
  [5] broadcast_unalias
    @ .\broadcast.jl:941
  [6] preprocess
    @ .\broadcast.jl:948
  [7] preprocess_args
    @ .\broadcast.jl:950
  [8] preprocess
    @ .\broadcast.jl:947
  [9] copyto!
    @ .\broadcast.jl:964
 [10] copyto!
    @ .\broadcast.jl:920
 [11] materialize!
    @ .\broadcast.jl:878
 [12] materialize!
    @ .\broadcast.jl:875
 [13] solve_equation
    @ .\REPL[5]:7

Stacktrace:
  [1] julia_error(msg::String, val::Ptr{…}, errtype::Enzyme.API.ErrorType, data::Ptr{…}, data2::Ptr{…}, B::Ptr{…})
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\errors.jl:384
  [2] julia_error(cstr::Cstring, val::Ptr{…}, errtype::Enzyme.API.ErrorType, data::Ptr{…}, data2::Ptr{…}, B::Ptr{…})
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\errors.jl:210
  [3] EnzymeCreatePrimalAndGradient(logic::Enzyme.Logic, todiff::LLVM.Function, retType::Enzyme.API.CDIFFE_TYPE, constant_args::Vector{…}, TA::Enzyme.TypeAnalysis, returnValue::Bool, dretUsed::Bool, mode::Enzyme.API.CDerivativeMode, runtimeActivity::Bool, width::Int64, additionalArg::Ptr{…}, forceAnonymousTape::Bool, typeInfo::Enzyme.FnTypeInfo, uncacheable_args::Vector{…}, augmented::Ptr{…}, atomicAdd::Bool)
    @ Enzyme.API C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\api.jl:268
  [4] enzyme!(job::GPUCompiler.CompilerJob{…}, mod::LLVM.Module, primalf::LLVM.Function, TT::Type, mode::Enzyme.API.CDerivativeMode, width::Int64, parallel::Bool, actualRetType::Type, wrap::Bool, modifiedBetween::NTuple{…} where N, returnPrimal::Bool, expectedTapeType::Type, loweredArgs::Set{…}, boxedArgs::Set{…})
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:1592
  [5] codegen(output::Symbol, job::GPUCompiler.CompilerJob{…}; libraries::Bool, deferred_codegen::Bool, optimize::Bool, toplevel::Bool, strip::Bool, validate::Bool, only_entry::Bool, parent_job::Nothing)
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:4436
  [6] codegen
    @ C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:3239 [inlined]
  [7] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:5288
  [8] _thunk
    @ C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:5288 [inlined]
  [9] cached_compilation
    @ C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:5340 [inlined]
 [10] thunkbase(mi::Core.MethodInstance, World::UInt64, FA::Type{…}, A::Type{…}, TT::Type, Mode::Enzyme.API.CDerivativeMode, width::Int64, ModifiedBetween::NTuple{…} where N, ReturnPrimal::Bool, ShadowInit::Bool, ABI::Type, ErrIfFuncWritten::Bool, RuntimeActivity::Bool, edges::Vector{…})
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:5451
 [11] thunk_generator(world::UInt64, source::LineNumberNode, FA::Type, A::Type, TT::Type, Mode::Enzyme.API.CDerivativeMode, Width::Int64, ModifiedBetween::NTuple{…} where N, ReturnPrimal::Bool, ShadowInit::Bool, ABI::Type, ErrIfFuncWritten::Bool, RuntimeActivity::Bool, self::Any, fakeworld::Any, fa::Type, a::Type, tt::Type, mode::Type, width::Type, modifiedbetween::Type, returnprimal::Type, shadowinit::Type, abi::Type, erriffuncwritten::Type, runtimeactivity::Type)
    @ Enzyme.Compiler C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\compiler.jl:5636
 [12] autodiff
    @ C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\Enzyme.jl:485 [inlined]
 [13] autodiff(::ReverseMode{…}, ::typeof(loss), ::Type{…}, ::Duplicated{…}, ::Const{…}, ::Duplicated{…})
    @ Enzyme C:\Users\xinle\.julia\packages\Enzyme\ydGh2\src\Enzyme.jl:524
 [14] top-level scope
    @ REPL[12]:1
Some type information was truncated. Use `show(err)` to see complete types.

This result was run with Enzyme v0.13.24 using Julia 1.10.2. The same error occurs on Enzyme v0.13.23 on Julia 1.11.1 as well. However, when the same code is tested on Enzyme v0.13.23 on Julia 1.10.1, the code runs successfully.