Hello, I am using ForwardDiff to solve a minimization problem. Following the suggestions I read, I need to specify what were my Float64 arrays as Real arrays. I do so and my code works! However, it is really slow compared to before. Basically, all I do is: M = zeros(Real,N, N,T); instead of M = …

I can be completely lost as I do not use this kinda of optimization, but using JuMP, Ipopt function myfun(x :: T, y :: T) where {T <: Real} X = zeros(T, 10,10); X .+= x; res1 = x^2 - sqrt(y); res2 = x + log(y); return maximum(abs.([res1;res2])); end model = Model(Ipopt.Optimize…

ForwardDiff - slowed down by Real vs. Float operations

General Usage

danicaratelli April 3, 2020, 9:51pm 6

Hi Sukera,

yes, the problem is with ForwardDiff, it seems to be a common problem (see here).

Best,

Topic		Replies	Views
Dual{::ForwardDiff...} type error when JuMP tries autodiff User-defined Function Optimization (Mathematical)	20	4554	April 28, 2021
ForwardDiff: Using on functions that are defined on floats64 General Usage	8	2528	December 2, 2018
ForwardDiff and Floats Optimization (Mathematical) forwarddiff	13	1047	April 28, 2021
Automatic differentation of `f(x::Real) = f(float(x))` leads to stackoverflow General Usage question	7	593	October 9, 2018
Type stability, deprecation Performance	4	839	November 8, 2017

ForwardDiff - slowed down by Real vs. Float operations

Related topics