EnsembleGPUArray() incompatibility with SecondOrderODEProblem()

Hi, I attempted to solve a second-order differential equation using a Julia library. However, I encountered an issue with the EnsembleGPUArray function, as the generated ensemble is not working as expected. Below, you can find the code that I used:

using DifferentialEquations, DataFrames, CSV, DelimitedFiles, RecursiveArrayTools, DiffEqGPU, CUDA

function lorenz_ddu(ddu, du, u, p, t)
    ddu[1] = p[1]*(du[2]-u[1])
    ddu[2] = u[1]*(p[2]-u[3]) - u[2]
    ddu[3] = u[1]*u[2] - p[3]*u[3]
end

function ensembleOutputExtracter(simulationResult, numTrajectories)
    for j = 1:min(numTrajectories, length(simulationResult))
        solnew = vcat(j, simulationResult[j].t', simulationResult[:, :, j])
        open("shahab.txt", "a+") do file
            writedlm(file, solnew', ',')
        end
    end
end

CUDA.set_runtime_version!("local")

u0 = [1.0, 1.0, 1.0]
du0 = [1.1, 1.2, 1.3]
tspan = (0.0, 100.0)
p = [10.0, 28.0, 8/3]
prob = SecondOrderODEProblem(lorenz_ddu, du0, u0, tspan, p)
prob_func = (prob, i, repeat) -> remake(prob, u0=rand(3).*u0, p=rand(3).*p)
monteprob = EnsembleProblem(prob, prob_func=prob_func, safetycopy=false)
sol = solve(monteprob, Tsit5(), EnsembleGPUArray(CUDA.CUDABackend(), 0.0), save_everystep=false, trajectories=10, saveat=1.0f0)
ensembleOutputExtracter(sol, 10)

Am I missing something?
Thanks,

Start by running the provided SecondOrderODEProblem and EnsembleGPUArray reproducer with the listed Julia packages and CUDA configuration. Trace the failure from the ensemble solve and compare the behavior with the supported problem types; the issue is done when the incompatibility is resolved or its supported behavior is clearly established.