Benoît Legat comments

Results 528 comments of


                                            Benoît Legat

Allow creating model from parts

I don't understand the codecov missing lines, it's reporting the signature of function as not covered but then its body is covered...

Differentiation of conic programs with nonlinear parametric expressions

I also faced the same issue, we should start by adding support for `ScalarNonlinearFunction` in ParametricOptInterface.

Convert common sub-functions as common sub-expressions

Consider the benchmark ```julia using JuMP f(x, u) = [sin(x[1]) - x[1] * u, cos(x[2]) + x[1] * u] function RK4(f, X, u) k1 = f(X , u) k2 =...

Convert common sub-functions as common sub-expressions

I fixed the performance issue of `check_belongs_to_model` and the issue of `map_indices` creating duplicates, it works now even without direct model ```julia using JuMP f(x, u) = [sin(x[1]) - x[1]...

Convert common sub-functions as common sub-expressions

I agree, it's not a clear choice. What convinced me to write this PR is the following. When users share sub-expressions by reference, we either: 1. Exponentially blow up the...

Convert common sub-functions as common sub-expressions

Do we have a benchmark showing that the subexpressions are slow ? I think you shared one at some point but I don't remember if it showed that it was...

Convert common sub-functions as common sub-expressions

This part should be enough for direct mode, the other parts were moved to https://github.com/jump-dev/MathOptInterface.jl/pull/2802 and https://github.com/jump-dev/MathOptInterface.jl/pull/2803

[Tracy] Update to v0.11.1

Thanks for the suggestion, let's try

Test ConstraintDual for bridges

Yes, we just check that get and set are inverse of each other and that's precisely what this PR is fixing. Now it also checks that it is the adjoint...

Fix bridge choice for non-Float64

Thanks for the reminder, I've just rebased, let's see what ci tells us