Intel 64 and IA-32 Architectures Software Developers Manual Volume 1, Basic Architecture

8-18 Vol. 1
PROGRAMMING WITH THE X87 FPU
8.2 X87 FPU DATA TYPES
The x87 FPU recognizes and operates on the following seven data types (see Figures
8-13): single-precision floating point, double-precision floating point, double
extended-precision floating point, signed word integer, signed doubleword integer,
signed quadword integer, and packed BCD decimal integers.
For detailed information about these data types, see Section 4.2.2, “Floating-Point
Data Types,” Section 4.2.1.2, “Signed Integers,” and Section 4.7, “BCD and Packed
BCD Integers.”
With the exception of the 80-bit double extended-precision floating-point format, all
of these data types exist in memory only. When they are loaded into x87 FPU data
registers, they are converted into double extended-precision floating-point format
and operated on in that format.
Denormal values are also supported in each of the floating-point types, as required
by IEEE Standard 754. When a denormal number in single-precision or double-preci-
sion floating-point format is used as a source operand and the denormal exception is
masked, the x87 FPU automatically normalizes the number when it is converted to
double extended-precision format.
When stored in memory, the least significant byte of an x87 FPU data-type value is
stored at the initial address specified for the value. Successive bytes from the value
are then stored in successively higher addresses in memory. The floating-point
instructions load and store memory operands using only the initial address of the
operand.