@certik Builtin % in LPython doesn't work on arrays by default because it is implemented as a function and not a binop. So the question is which approach to follow,

1. To make _mod function defined in lpython_builtin.py vectorised;
2. Define Mod as binop and then use builder->CreateSRem in LLVM backend. Note that not all backends might support modulus operation on types other than integers.
3. Define mod and then make it vectorisable.

I think 3rd approach is a good one and complies with Python and NumPy standards. What do you say @certik?

Originally posted by @czgdp1807 in #1002 (comment)