LBFGS: Does Optim.jl also implement an improvement over the original routine (which scipy.optimize does)?

[rht]

@IlyaOrson said on Julia Discourse (https://discourse.julialang.org/t/optim-jl-vs-scipy-optimize-once-again/61661/25):

Does Optim.jl also implement the newer modifications referenced at the start of the fortran code used by scipy? That paper is not referenced by Optim

All the modifications in that lbfgsb.f are preceded by a line comment containing c-jlm-jn.

[rht]

After the discussion in Gnimuc/LBFGSB.jl#11, I found out that LBFGSB.jl uses the code that has the c-jlm-jn lines (2011 version). But since this is L-BFGS-B, that Optim.jl doesn't have yet. It makes the issue irrelevant. And in the future when Optim.jl does contain the L-BFGS-B code, it will automatically have the c-jlm-jn modifications.