[next] [prev] [prev-tail] [tail] [up]

4.2 IEEE floating point formats

The IEEE (Institute of Electrical and Electronics Engineers) has established two floating point number formats. A single precision format (float in C) uses 32 bits, while a double precision format (double in C) uses 64 bits.

Counting bits from the least significant digit, table 1 describes the two formats.

Format SignExponentExp. Bias,Mantissa
Single precision 31 23-30 127 0-22
Double precision64 52-62 1023 0-51

Table 1: IEEE floating point formats.
  4.2.1 Special values
  4.2.2 Example 1
  4.2.3 Example 2
  4.2.4 Visualizing floating point representation

[next] [prev] [prev-tail] [front] [up]