6. Floating instructions

The various instruction sets feature floating-point instructions for single-precision (32-bit) and double-precision (64-bit) operations. This is the case for all general-purpose processors and high-end processors used in signal processing.

General registers are defined for floating-point data. Most instruction sets feature basic single- and double-precision arithmetic operations, including absolute value, addition, subtraction, multiplication, division and square root, memory instructions for loading and storing floating-point data, and comparison instructions for positioning Booleans in the same way as comparison instructions for integer data.

For reasons linked to the way in which floating-point instructions were implemented in a particular coprocessor called 8087, the specification of floating-point instructions in IA-32 is particular....