Here we are: https://gitlab.com/alanmarazzi/numpy-vs-neanderthal/blob/master/README.md

I included Numpy results with both eig and eigh (that's the only difference), Neanderthal is almost always faster and I would say more reliable (less variance)

nice, the numbers look promising

how do neanderthal and numpy compare feature wise?

@chrisblom Neanderthal has a subset of Numpy functions (though @blueberry was adding some for the next releases), and then has some more advanced methods that if I understand correctly are material for a library built on top of Neanderthal

To better understand: calculating covariance (Numpy's cov function) will probably never be implemented in Neanderthal

ok, so it is more about bare-bones matrix manipulation

and higher level operations will be in a lib on top of that

Well, "will be" is a bit strong, someone would have to code that 😅

some experiments of using R from clojurescript: https://www.maria.cloud/gist/6a0b78b82f52f5b1bff64053a00660dd -- would love to hear your comments! worth mentioning here: @mhuebert, for the maria.cloud tool @carsten.behring, for the idea of using opencpu @jonyepsilon, for the gg4clj syntax