diffeqpy icon indicating copy to clipboard operation
diffeqpy copied to clipboard

could not load library "libopenlibm"

Open zmoon opened this issue 3 years ago • 3 comments

    sol = de.solve(prob, solver)
RuntimeError: <PyCall.jlwrap (in a Julia function called from Python)
JULIA: could not load library "libopenlibm"
The specified module could not be found.
Stacktrace:
 [1] log10 at .\math.jl:566 [inlined]
 [2] ode_determine_initdt(::Array{Float64,2}, ::Float64, ::Float64, ::Float64, ::Float64, ::Float64, ::typeof(DiffEqBase.ODE_DEFAULT_NORM), ::ODEProblem{Array{Float64,2},Tuple{Float64,Float64},true,TendParams,ODEFunction{true,typeof(tend!),LinearAlgebra.UniformScaling{Bool},...
...

I couldn't find info about how to fix this. I don't have this issue when Python isn't involved (using DifferentialEquations.jl in Julia normally). Let me know if you want to me to provide more info.

zmoon avatar Mar 01 '21 20:03 zmoon

Is this using the standard generic binary?

ChrisRackauckas avatar Mar 01 '21 21:03 ChrisRackauckas

Yes, though I also tried using a sysimage for DifferentialEquations that I created with PackageCompiler and got the same error.

This is on Windows btw. I did get it to work on Linux (WSL) on the same machine after doing an --enable-shared Python build.

zmoon avatar Mar 01 '21 21:03 zmoon

This might need to be an issue in pyjulia.

ChrisRackauckas avatar Mar 01 '21 22:03 ChrisRackauckas

Fixed by PythonCall

ChrisRackauckas avatar Oct 20 '23 02:10 ChrisRackauckas