`realmin`

Smallest positive normalized floating-point number.

Category: Language Fundamentals > Data Types

Syntax

realmin

Smallest normalized double, about 2.2251e-308.

realmin('single')

Smallest normalized single, about 1.1755e-38.

Description

Below realmin numbers become denormalized: they lose precision gradually until they underflow to zero. realmin is therefore the smallest value that keeps the full precision of the type. Do not confuse it with eps, which measures the spacing of numbers near 1.

Input arguments

classname: 'double' (default) or 'single'.

Output arguments

x: Scalar of the requested class.

Example

realmin
realmin / 2 > 0
realmin('single')

Output

ans = 2.2251e-308
ans = logical
   1
ans = single
   1.1755e-38

realmin