First page Back Continue Last page Overview Graphics
Floats, Cont'd
Currently, most manufacturers adopt IEEE standard:
- s = 1st bit
- Next 8 bits are e
- Last 23 bits are M, expressed as a binary fraction, either 1.F, or, if e=0, 0.F, where F is in base 2.
Largest single-precision float fmax = 2127 1038.
Smallest (and least precise!) fmin = 2-149 10-45.
Notes: