Investigate the variations in multiply and divide instructions supported by three different microprocessor families. Why is there more variation in the ways in which multiplication and division are performed than in, for example, addition?

Question

What will be an ideal response?

Answer 1

The first part of this question requires students to look up multiplication instructions from three different
computers.

The reason that there is very little variation in addition operations is that adding two m?bit words yields an mbit result and a carry out. Some computers have a variation in which a carry?in from a previous operation is included. Moreover, the addition of signed 2‘s complement values is the same as the addition of unsigned values. Consequently, you don’t need separate addition instructions for signed and unsigned operations.

Multiplication is more complex because the multiplication of two m?bit values yields a 2m?bit product. You then have to decide whether to keep the upper or lower halves of the result if you are using m?bit arithmetic. If you are using 2m?bit arithmetic you have to put the result in two registers. Finally, multiplication does not give the same value for signed and unsigned numbers, which means that you have to have signed and unsigned multiplication operations.

Division is like multiplication in the sense that not all operands are the same size. Moreover, division yields a quotient and a remainder. When defining a division instruction you have (a) length of divisor, (b) length of dividend, (c) length of quotient, and (d) length of remainder. A typical approach is to implement the division of a 32?bit value by a 16?bit value giving a 16?bit quotient and a 16?bit register. Note, that the quotient and remainder (in the case of the 68K) are packed into the same register.