# Module Float

- Constant
**DIGITS_10**

Constant**MIN_10_EXP**

Constant**MAX_10_EXP**

Constant**MIN**

Constant**MAX**

Constant**EPSILON**

`constant`

Float.`DIGITS_10`

`constant`

Float.`MIN_10_EXP`

`constant`

Float.`MAX_10_EXP`

`constant`

Float.`MIN`

`constant`

Float.`MAX`

`constant`

Float.`EPSILON`

- Description
These constants define the limits for floats on the current architecture:

- DIGITS_10
The number of decimal digits that can be represented. Any number with this many decimal digits can be stored in a float and converted back to decimal form without change. DIGITS_10 is not less than

`6`

.- MIN_10_EXP
- MAX_10_EXP
Limits of the exponent in decimal base. 10 raised to any number within this range can be represented in normalized form. MIN_10_EXP is not greater than

`-37`

. MAX_10_EXP is not less than`37`

.- MIN
The smallest normalized float greater than zero. It's not greater than

`1e-37`

.- MAX
The largest finite float. It's not less than

`1e37`

.- EPSILON
The difference between 1 and the smallest value greater than 1 that can be represented. It's not greater than

`1e-5`

.

- Note
The size of the float type can be controlled when Pike is compiled with the configure flags

`--with-double-precision`

and`--with-long-double-precision`

. The default is to use the longest available float type that fits inside a pointer.

- Constant
**FLOAT_PRECISION**

Constant**DOUBLE_PRECISION**

Constant**LONG_DOUBLE_PRECISION**

`constant`

Float.`FLOAT_PRECISION`

`constant`

Float.`DOUBLE_PRECISION`

`constant`

Float.`LONG_DOUBLE_PRECISION`

- Description
Tells which C compiler float type that is used for Pike floats. Only one of these constants will exist (with the value

`1`

) at runtime.- FLOAT_PRECISION
The

`float`

type of the C compiler is used.- DOUBLE_PRECISION
The

`double`

type of the C compiler is used.- LONG_DOUBLE_PRECISION
The

`long double`

type of the C compiler is used.

- Note
The float type can be controlled when Pike is compiled with the configure flags

`--with-double-precision`

and`--with-long-double-precision`

. The default is to use the longest available float type that fits inside a pointer.