Yingbo Ma

> Which platform are you using? I am using Arch Linux. > Did you invoke Julia using julia -i ... ? No, I didn't.

Unfortunately, it doesn't seem to fix the problem for me ```julia ➜ env GKS_NO_EXIT_HANDLER=1 julia _ _ _ _(_)_ | Documentation: https://docs.julialang.org (_) | (_) (_) | _ _ _|...

They should both work.

Hey, could you run the above benchmark again using the latest MTK? It should be able to handle larger systems now.

Thanks for taking the time to run it again. For large and complex systems, symbolically computing the Jacobian won't be a good idea because the size of the expression will...

Also, we are now defaulting to KLU for sparse matrices with moderate size, so you can just use `QBDF(autodiff=false)`, `QNDF(autodiff=false)`, or `FBDF(autodiff=false)`.

BTW, the warning is fixed in https://github.com/SciML/ModelingToolkit.jl/pull/1895

I have modified the script a bit to make it friendlier to the compiler. Could you try this as well? ```julia using ModelingToolkit, OrdinaryDiffEq, Symbolics, IfElse, XSteam, Polynomials, Plots, LinearSolve...

I started Julia with O0, and with N=2000, I got ```julia julia> sol = @time solve(prob, solver, reltol=1e-6, abstol=1e-6, saveat=100); 3923.945535 seconds (363.26 M allocations: 99.133 GiB, 54.01% gc time,...

You probably don't want to use Rosenbrock methods for this. I will look into it.