Performs two-way pack and unpack for IEEE 754 floating-point numbers
represented as words.
Not designed for performance. Uses CLP(FD) for bit manipulation. and
hence remains within the integer domain. Float arithmetic applies
outside the finite-domain constraints.
- Arguments:
-
Word | - is a non-negative integer. This implementation does not
handle negative integers. Negative support implies a non-determinate
solution for packing. A positive and negative answer exists for any
given Float. |
Sig | - is the floating-point significand between plus and minus 1.
Uses Sig rather than Mantissa; Sig short for Significand, another
word for mantissa. |