warp.argmax
===========

.. function:: warp._src.lang.argmax(a: Vector[Scalar, Any]) -> uint32

   .. hlist::
      :columns: 8

      * Kernel

   Compute the index of the maximum element of vector ``a``.
   
   On float types, NaN elements are skipped; the result is the index of the
   largest non-NaN element. If every element is NaN, returns ``0``.