fix16_mul, fix16_div, fix16_sqrt. They are faster & more accurate than the previous versions. Closes issue #13. Includes unittests for the functions in question, runnable by typing 'make' in the unittests folder.