Inside Macintosh: PowerPC Numerics / Part 2 - The PowerPC Numerics C Implementation

Chapter 7 - Numeric Data Types in C

This chapter describes the numeric data types available in C and shows how to determine the class and sign of values represented in numeric data types. As stated in Chapter 2, "Floating-Point Data Formats," the PowerPC Numerics environment provides three numeric data formats: single (32 bits long), double (64 bits long), and double-double (two double formats combined, resulting in 128 bits). Each can represent normalized numbers, denormalized numbers, zeros, NaNs, and Infinities. See Chapter 2 for information about the numeric data formats and about how they represent values. Read this chapter to find out about the mapping of numeric formats to floating-point types in C, about the floating-point type declarations made in the PowerPC Numerics library (MathLib), and about the library utilities available that can determine the class of a floating-point value.

---

Chapter Contents

C Data Types

Efficient Type Declarations

Inquiries: Class and Sign

Creating Infinities and NaNs

Numeric Data Types Summary

C Summary

Constants

Class and Sign Inquiry Macros

Data Types

Special Value Routines

Creating NaNs