I know ForwardDiff no longer exports the derivative functions, so someone will have to go in and change those kinds of things.
Paweł Biernat
@pwl
I don't know how @PerezHz would feel about it, maybe you should write him first?
Christopher Rackauckas
@ChrisRackauckas
Okay, I'll email him.
Thanks for the heads up on this
Paweł Biernat
@pwl
if you won't be able to reach him you can try to write @lbenet, he will know what the status of the package is
I think he is his advisor, but I'm not so sure about it now:-)
Jorge Pérez
@PerezHz
Hi everyone! @ChrisRackauckas thanks for contacting us, glad to know you're interested in what we've been working on! :smile: Yes, we did take it down, we're re-organizing it (it's pretty much mixed with some celestial mechanics stuff) and planning to re-upload the generic parts of the code
Christopher Rackauckas
@ChrisRackauckas
Okay cool!
I am currently wrapping a bunch of packages into DifferentialEquations.jl, and would like to have that as one of the options to choose from.
Just let me know when you got something you want to show.
Are there any plans to release the interval arithmetic ODE solver? I saw something like that on @lbenet presentation.
Jorge Pérez
@PerezHz
Sure Chris, we'll let you know!
Yes @pwl there are plans, we're still working on it; the thing is, the leap towards interval arithmetics is non-trivial enough, @lbenet can give you more details about that
Christopher Rackauckas
@ChrisRackauckas
Are you going to be using ArbReals? I know dpsanders is in on that project.
This application would be perfect for ArbReals though.
Though Interval{Float64} would be faster for small calculations, Interval{BigFloat} would do better than ArbReals if you need 500+ bits of accuracy.
But in the middle, 64-~500 bits, ArbReals is the way to go.
I kow Sanders and Benet have been on the emails with Jeffrey Sarnoff on ArbFloats/ArbReals, so I wonder if that's why.
Felix Henneke
@fhenneke
I might be mistaken, but should the name fehlberg in runge-kutta-fehlberg not be spelled with an E instead of a U?
Christopher Rackauckas
@ChrisRackauckas
Yes.
Where's do I have the error?
Felix Henneke
@fhenneke
everywhere? e.g. on the first page of the documentation and in the ode_tableaus.jl in the source code
Christopher Rackauckas
@ChrisRackauckas
oh...
:worried:
Thanks for letting me know!
Felix Henneke
@fhenneke
btw: keep up the good work! :)
Christopher Rackauckas
@ChrisRackauckas
Thanks!
Romain Veltz
@rveltz
HI,
I tried the simple code 'using DifferentialEquations
function vanDerPolExample(u₀=[0,sqrt(3)]) f1(u,t) = (1-u[2].^2)*u[1] - u[2] f2(u,t) = u[1] f(u,t) = [f1(u,t);f2(u,t)] return(ODEProblem(f,u₀)) end prob = vanDerPolExample() Δt = 1//2^(4) #The initial timestepping size. It will automatically assigned if not given. tspan = [0,20] # The timespan. This is the default if not given. sol = solve(prob::ODEProblem,tspan,Δt=Δt,alg=:RK4)'
but it gives the error (line 12 is the solve):
ERROR: LoadError: BoundsError in getindex at ./number.jl:21 [inlined] in (::#f1#10)(::Rational{Int64}, ::Array{Float64,1}) at /Users/rveltz/Downloads/ode.jl:4 in (::#f#12{#f1#10,#f2#11})(::Rational{Int64}, ::Array{Float64,1}) at /Users/rveltz/Downloads/ode.jl:6 in (::DifferentialEquations.##500#507{DifferentialEquations.ODEProblem{Array{Float64,1},Float64}})(::Rational{Int64}, ::Array{Float64,1}, ::Array{Float64,1}) at /Users/rveltz/.julia/v0.5/DifferentialEquations/src/ode/ode_solve.jl:79 in ode_solve(::DifferentialEquations.ODEIntegrator{:RK4,Array{Float64,1},Float64,2,Rational{Int64}}) at /Users/rveltz/.julia/v0.5/DifferentialEquations/src/ode/ode_integrators.jl:376 in #solve#499(::Array{Any,1}, ::Function, ::DifferentialEquations.ODEProblem{Array{Float64,1},Float64}, ::Array{Int64,1}) at /Users/rveltz/.julia/v0.5/DifferentialEquations/src/ode/ode_solve.jl:134 in (::DifferentialEquations.#kw##solve)(::Array{Any,1}, ::DifferentialEquations.#solve, ::DifferentialEquations.ODEProblem{Array{Float64,1},Float64}, ::Array{Int64,1}) at ./<missing>:0 in include_from_node1(::String) at ./loading.jl:426 in include_from_node1(::String) at /Applications/Julia-0.5.app/Contents/Resources/julia/lib/julia/sys.dylib:? while loading /Users/rveltz/Downloads/ode.jl, in expression starting on line 12
Do you have any idea?
Thank you,
Christopher Rackauckas
@ChrisRackauckas
Are you on master? Then the API changed to (t,u) and (t,u,du) to match ODE.jl, ODEInterface.jl, and Sundials.jl. IF you're not on master, than it's okay.
Christopher Rackauckas
@ChrisRackauckas
You may want to look at the updated docs if you're on master (/ the new version is coming out). That's pretty much the only user facing change.
Christopher Rackauckas
@ChrisRackauckas
But that's gotta be it: the error happened on 376 of ode_integrators which is just f(t,u,k₁), which is just k₁[:]=f(t,u), so the error is inside the function call, and it's likely erroring because if u and t are flip then there's no t[1].
Romain Veltz
@rveltz
That was it! The inplace modification is still available then?
Christopher Rackauckas
@ChrisRackauckas
Yes. It's (t,u,du). Julia usually has the in-place moded variable first, but all the Fortran codes that we have wrapped have them last. So it's slightly more efficient for the Fortran codes to just have everything be (t,u,du). ODE.jl for some reason already had that ordering, so that means all DE packages (that I know of) are standardized on it now.