Inside Macintosh: PowerPC Numerics / Part 1 - The PowerPC Numerics Environment

Chapter 2 - Floating-Point Data Formats

This chapter describes the data formats your PowerPC application can use to represent floating-point numbers. It begins by discussing in general the methods PowerPC Numerics uses to store and interpret floating-point values and by explaining why those methods were chosen. The chapter introduces the special values zero, NaN (Not-a-Number), and Infinity and explains why these special values are necessary. Next is an in-depth description of the numeric data formats with a discussion of how these formats represent floating-point values. At the end of the chapter, you will find a table comparing the size, range, and precision of the numeric data formats. This table can help you choose which data format is best for your application.

You should read this chapter to learn about the floating-point data formats available on PowerPC processor-based Macintosh computers and to learn more about how your computer encodes and manipulates floating-point numbers.

---

Chapter Contents

About Floating-Point Data Formats

Interpreting Floating-Point Values

Normalized Numbers

Denormalized Numbers

Infinities

NaNs

Zeros

Formats

Single Format

Double Format

Double-Double Format

Range and Precision of Data Formats