X86 Assembly Language and C Fundamentals

Chapter 11

Floating-Point Arithmetic Instructions

The fixed-point number representation is appropriate for representing numbers with small numerical values that are considered as positive or negative integers; that is, the implied radix point is to the right of the low-order bit. The same algorithms for arithmetic operations can be employed if the implied radix point is to the immediate right of the sign bit, thus representing a signed fraction.

The range for a 16-bit fixed-point number is from (–215) to (+215 – 1), which is inadequate for some numbers; for example, the following operation:

$28,400,000,000.\times 0.0000000546$

This operation can also be written in scientific notation, as follows:

$(0.284\times {10}^{11})\times (0.546\times {10}^{-7})$

where 10 is the base and 11 ...