In the old days, floating-point math required using either software emulation or a separate math co-processor. Since the 80486, Intel has incorporated the floating-point operations in an onboard FPU (see Chapter 2, "The IA-32 Platform"). This chapter describes the floating-point operations contained within the FPU and demonstrates how to perform floating-point math on the IA-32 platform.

The first part of this chapter describes the layout of the FPU, and recaps the instructions demonstrated in Chapter 7, "Using Numbers," for loading numbers into the FPU and retrieving results from the FPU. Next, the basic floating-point math functions are shown: addition, subtraction, multiplication, and division. Following that, you will learn how to work with the more advanced floating-point math functions, such as square roots and trigonometric functions. After that, the methods used for comparing floating-point numbers are described, followed by the methods used for storing the FPU environment in memory, and restoring the FPU environment from the backup.

Chapter 2 described the basics of the FPU environment on the IA-32 platform. Now that you are more familiar with the layout and operation of the IA-32 platform, it's time to dig a little deeper and examine both the FPU infrastructure and the instructions that are used to control it. This section describes the FPU register stack: the control word, which controls how the FPU operates; the status ...

Start Free Trial

No credit card required