1. 6
  1.  

  2. 1

    Sometimes you don’t have much control over the data type. For example, Python uses doubles because their overhead is negligible when compared with the interpreter. If you want more control (and speed) you need to use NumPy.