We learned that MIPS provides signed and unsigned addition and that some processors provide signed and unsigned multiplication. In this chapter, we encountered saturating arithmetic. It does seem rather sensible to tag data or instructions with the type of operand. What would be the advantages, disadvantages, and implications of such a step?

What will be an ideal response?

Tagging data means that appropriate operations can automatically be applied to data. This is, of course, rather
like strong typing in high level languages. By typing data you automatically perform the appropriate action; for
example, you use the same ADD for signed, unsigned and saturating type data. The advantage of typing is that
you need fewer instructions. The disadvantage is that the data word must be extended to include tags; for
example, an additional two bits would be required to specify four types. The Itanium IA64 has a 65?bit register
with 64 bits data and a NaT, not a thing tag, to indicate the data value is not valid. This is used by shortcircuiting
calculations; all calculations involving a NaT yield a NaT result.