Liang Wang

It is a shame ... but you are right. It has not been implemented yet. I had a plan long time ago about building a statistical learning module/toolbox in owl,...

@jzstark is this still relevant? or implemented already?

Very insightful discussions :+1:

@jzstark can you test on both the mac and thinkpad. It seems I have some compilation issue on one of my thinkpads. I will also dig a bit deeper.

> I tried compiling the code with this PR on my Ubuntu 20.04 desktop, and there are indeed errors and warnings; the current master branch compile normally. Part of the...

> -Wno-logical-op-parentheses From here https://github.com/owlbarn/owl/blob/eead31f1920e63b876421e6a30e61c947f29827f/src/owl/config/configure.ml#L115 Actually, we can simply remove get_devmode_cflags, and move [ "-Wall"; "-pedantic"; "-Wextra"; "-Wunused" ] to https://github.com/owlbarn/owl/blob/eead31f1920e63b876421e6a30e61c947f29827f/src/owl/config/configure.ml#L253

@murbard oh sure, this has been on my TODO for some time. Einstein sum will be added.

Dhruv, thanks for pointing it out. That is a bug needs to be fixed.

This benchmark looks cool! I am (kinda) aware of this but nobody really did any benchmarking before to show how slow it is, so really appreciate this. Lacaml uses LAPACK...

Acturally, it would be very interesting to know -- how the perfermance difference varies as a function of matrix size ... probably at some point, both lacaml and owl reach...