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

Chapter 5 - Conversions

Converting Integers to Floating-Point Formats

When an integer is converted to a floating-point format whose precision is greater than or equal to the size of the integer format, the conversion is exact. When an integer is converted to a floating-point format whose precision is less than the size of the integer format, the integer is rounded in the current rounding direction. For example, because the single format has 24 bits in the significand, any integer requiring more than 24 bits of precision will not be converted to its exact value.