Hi @heltonmc, I think everyone appreciates your positive tone in the discussion.
I resisted the temptation for several days to participate in the discussions of this thread any further because I do not see where it is heading. Such discussions can become more objective if people pay more attention to a few details.
-
Python, SciPy, or any other wrapper is not Fortran. Please clarify to the Julia community that these three are not the same if you can. Testing Julia packages against Python wrappers that may or may not call some dynamic C/Fortran libraries is not the same as benchmarking it against pure C/Fortran. Please look at the link you posted in your comment and see how the accepted answer uses MATLAB, SciPy, and Fortran effectively synonymously.
I remember @oscardssmith 's amazement in this discourse at the fact that Fortran compilers can indeed unroll loops. As far as I remember from his comment, someone had shown him a Python-Fortran wrapper benchmark in JuliaCon19 that somehow proved Fortran compilers incapable of loop-unrolling (I vaguely remember the discussion, apologies in advance if it is incomplete or inaccurate). -
When someone from a corporation comments on a topic in this forum, they always begin the discussion by introducing themselves, mentioning their affiliations, and revealing any conflicts of interest. It would certainly help if people from Julia Discourse and JuliaComputing followed suit in this forum to eliminate irrelevant and unproductive comments (and sometimes accusations) from this forum.