Return of ODESolution type fails

[ChrisRackauckas]

With https://github.com/JuliaDiffEq/DiffEqBase.jl/pull/240 and SimpleDiffEq#master, the entire returned object is immutable. However, it segfaults when it returns. But, returning it's solution array, sol.u, is fine.

using GPUifyLoops, CuArrays, SimpleDiffEq
using StaticArrays, Cthulhu

function loop(u, p, t)
    @inbounds begin
        σ = p[1]; ρ = p[2]; β = p[3]
        du1 = σ*(u[2]-u[1])
        du2 = u[1]*(ρ-u[3]) - u[2]
        du3 = u[1]*u[2] - β*u[3]
        return SVector{3}(du1, du2, du3)
    end
end
function liip(du, u, p, t)
    σ = p[1]; ρ = p[2]; β = p[3]
    du[1] = σ*(u[2]-u[1])
    du[2] = u[1]*(ρ-u[3]) - u[2]
    du[3] = u[1]*u[2] - β*u[3]
    return nothing
end

u0 = 10ones(Float32,3)
const su0= SVector{3}(u0)
const dt = 1f-1

odeoop = ODEProblem{false}(loop, SVector{3}(u0), (0.0f0, 10.0f0),  Float32[10, 28, 8/3])
sol2 = solve(odeoop,GPUSimpleTsit5(),dt=dt)
ps = CuArray([@SVector [10f0,28f0,8/3f0] for i in 1:10])
CuArrays.allowscalar(false)

function f(p)
    prob = ODEProblem{false}(loop, su0, (0.0f0, 10.0f0),  p)
    solve(prob,GPUSimpleTsit5(),dt=dt).u
end

_f = GPUifyLoops.contextualize(f)
map(_f,ps)

function f2(p)
    prob = ODEProblem{false}(loop, su0, (0.0f0, 10.0f0),  p)
    solve(prob,GPUSimpleTsit5(),dt=dt)
end

_f2 = GPUifyLoops.contextualize(f2)
map(_f2,ps)

[vchuravy]

Maybe I am on the wrong branch, but I am getting

julia> sol2 = solve(odeoop,GPUSimpleTsit5(),dt=dt)
ERROR: MethodError: Cannot `convert` an object of type Nothing to an object of type DiffEqBase.DEStats
Closest candidates are:
  convert(::Type{T}, ::T) where T at essentials.jl:154
  DiffEqBase.DEStats(::Any) at /home/vchuravy/.julia/packages/DiffEqBase/6ewdP/src/destats.jl:15
  DiffEqBase.DEStats(::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any) at /home/vchuravy/.julia/packages/DiffEqBase/6ewdP/src/destats.jl:2
Stacktrace:
 [1] ODESolution{Float32,2,SArray{Tuple{101},SArray{Tuple{3},Float32,1,3},1,101},Nothing,Nothing,StepRangeLen{Float32,Float64,Float64},Nothing,ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem},GPUSimpleTsit5,DiffEqBase.LinearInterpolation{StepRangeLen{Float32,Float64,Float64},SArray{Tuple{101},SArray{Tuple{3},Float32,1,3},1,101}}}(::SArray{Tuple{101},SArray{Tuple{3},Float32,1,3},1,101}, ::Nothing, ::Nothing, ::StepRangeLen{Float32,Float64,Float64}, ::Nothing, ::ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem}, ::GPUSimpleTsit5, ::Function, ::Bool, ::Int64, ::Nothing, ::Symbol) at /home/vchuravy/.julia/packages/DiffEqBase/6ewdP/src/solutions/ode_solutions.jl:2
 [2] #build_solution#14(::Bool, ::Bool, ::Bool, ::Bool, ::Nothing, ::Function, ::Symbol, ::Nothing, ::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::typeof(DiffEqBase.build_solution), ::ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem}, ::GPUSimpleTsit5, ::StepRangeLen{Float32,Float64,Float64}, ::SArray{Tuple{101},SArray{Tuple{3},Float32,1,3},1,101}) at /home/vchuravy/.julia/packages/DiffEqBase/6ewdP/src/solutions/ode_solutions.jl:57
 [3] (::getfield(DiffEqBase, Symbol("#kw##build_solution")))(::NamedTuple{(:k, :destats, :calculate_error),Tuple{Nothing,Nothing,Bool}}, ::typeof(DiffEqBase.build_solution), ::ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem}, ::GPUSimpleTsit5, ::StepRangeLen{Float32,Float64,Float64}, ::SArray{Tuple{101},SArray{Tuple{3},Float32,1,3},1,101}) at ./none:0
 [4] #solve#25(::Float32, ::Float32, ::Float32, ::Function, ::ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem}, ::GPUSimpleTsit5) at /home/vchuravy/.julia/packages/SimpleDiffEq/FTIOd/src/tsit5/gpuatsit5.jl:51
 [5] (::getfield(DiffEqBase, Symbol("#kw##solve")))(::NamedTuple{(:dt,),Tuple{Float32}}, ::typeof(solve), ::ODEProblem{SArray{Tuple{3},Float32,1,3},Tuple{Float32,Float32},false,Array{Float32,1},ODEFunction{false,typeof(loop),LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Nothing,DiffEqBase.StandardODEProblem}, ::GPUSimpleTsit5) at ./none:0
 [6] top-level scope at none:0

ah yes wrong branch

[vchuravy]

So investigating with Cthulhu:

• %108  = invoke overdub(::Cassette.Context{…},::typeof(DiffEqBase.__has_syms){…},::typeof(loop){…})::Any

which calls fieldnames, which is not an inferrable function