Liang Wang

ok, i can add ... another note, you may find this function useful as well https://github.com/owlbarn/owl/blob/5d652b10add0f9c31395e358f18b5180a8a7ecec/src/base/misc/owl_utils_array.mli#L347

sylvester fails randomly. Does it use randomly generated inputs? If so, one temporary solution is using fixed inputs.

This has been proposed before https://github.com/owlbarn/owl/issues/374 I need to see an exmaple to demonstrate the benefits and how it can be applied to owl and how much it will impact...

It was removed because the generated matrix still uses float numbers. Generating an integer matrix can be done in several ways, by casting functions, or round, floor, ceil, trunc, fix,...

About the inconsistency between dense and sparse is, long story short, I have been trying very hard to make dense and sparse have the same interface since the very beginning...

Good comment, really appreciate it. Indeed, interoperation on different number types is really desirable when fast prototyping. Most of the time in my research, I simply want to try something...

Probably the tests can be divided into several groups. - Basic operations like adding, removing, modifying nodes. - Shape inference functions. - Structural optimisation functions. - Anything esle? open to...

ordering is not required for `roots` afaik

Similar discussion found by @jzstark http://numpy-discussion.10968.n7.nabble.com/Silent-Broadcasting-considered-harmful-td39665.html Numpy decided not to change, the reason is that broadcast logic is implmented down to the c level for efficiency considerations, same in owl....

@bluddy thank you very much for the detailed explanation. I see your point and I think they are reasonable. Since the suggest change is quite significant so I would like...