Principles of Operation: toFloat converter
Aaron Stewart, 24 July 2001

This specific implementation works by casting the integer as
a floating-point value and then normalizing the value.

For example (using base-ten):

fixed-point input = 00015493
	(The leading zeros are included because the number of digits is fixed.)

= 0.0015493 x 10^7
	(The "7" is known at hardware construction time.)

This value is normalized by shifting left, and the amount of shift
is recorded.  In this implementation, a log shifter is used.
= 1.5493000 x 10^(7+3)

The normalized value goes into the floating-point significand.
The hard-coded exponent (7) and the run-time exponent (3) are
added to the exponent bias.  This is the floating-point exponent.

If the original input value is signed, then the input value is
converted to signed-magnitude form and the conversion proceeds
as usual.

-end-
