Important: The information in this document is obsolete and should not be used for new development.
Range and Precision of Data Formats
Table 2-7 shows the precision, range, and memory usage for each numeric data format. You can use this table to compare the data formats and choose which one is needed for your application. Typically, choosing a data format requires that you determine the tradeoffs between
In the table, decimal ranges are expressed as rounded, two-digit decimal representations of the exact binary values. The speed of a given data format varies depending on the particular implementation of PowerPC Numerics. (See Chapter 5, "Conversions," for information on aspects of conversion relating to precision.)
- fixed-point or floating-point form
- precision
- range
- memory usage
- speed
For example, in single format, the largest representable number is composed as follows:
The smallest positive normalized number representable in single format is made up as follows:
- significand
- .111111111111111111111112
- exponent
- value
- \xC5
For denormalized numbers, the smallest positive value representable in the single format is made up as follows:
- significand
- .000000000000000000000002
- exponent
- value
- \xC5
- significand
- .000000000000000000000012
- exponent
- value
- \xC5