I have to develop a routine for the 16c84 in which 2 three-byte
floating-point numbers were multiplied.There are no further
restrictions.What's the best way?.