A double value is 64 bits wide. Figure 2 shows its structure.
As before, S is the sign, Exp the
exponent, and Frac the fraction. Most of the
detail of float values remains true for double values,
except that:
The
Expfield is biased by0x3FF(1023) instead of0x7F, so numbers between 1.0 and 2.0 have anExpfield of0x3FF.The
Expvalue used to represent infinity and NaNs is0x7FF(2047) instead of0xFF.
See also- Concepts
- Reference
- Other information
IEEE Standard for Floating-Point Arithmetic (IEEE 754), 1985 version
PDF version