Optimized 2D-by-1D broadcasting in optimized op_mul (#4965)

Detect that we are doing an elementwise multiplication for a 2D tensor and a 1D tensor. Dispatch to a vectorized kernel for this case.

Differential Revision: [D61560826](https://our.internmc.facebook.com/intern/diff/D61560826/)

Pull Request resolved: https://github.com/pytorch/executorch/pull/4808

---------

Co-authored-by: Scott Wolchok <swolchok@fb.com>