The C Book

Two header files <float.h> and <limits.h> define several implementation specific limits.

9.5.1. Limits.h

Table 9.1 gives the names declared, the allowable values, and a comment on what they mean. For example, the description of SHRT_MIN shows that in a given implementation the value must be less than or equal to −32767: this means that for maximum portability a program cannot rely on short variables being able to hold values more negative than −32767. Implementations may choose to support values which are more negative but must provide support for at least −32767.

Name	Allowable value	Comment
`CHAR_BIT`	(≥8)	bits in a `char`
`CHAR_MAX`	see note	max value of a `char`
`CHAR_MIN`	see note	min value of a `char`
`INT_MAX`	(≥+32767)	max value of an `int`
`INT_MIN`	(≤−32767)	min value of an `int`
`LONG_MAX`	(≥+2147483647)	max value of a `long`
`LONG_MIN`	(≤−2147483647)	min value of a `long`
`MB_LEN_MAX`	(≥1)	max number of bytes in a multibyte character
`SCHAR_MAX`	(≥+127)	max value of a `signed char`
`SCHAR_MIN`	(≤−127)	min value of a `signed char`
`SHRT_MAX`	(≥+32767)	max value of a `short`
`SHRT_MIN`	(≤−32767)	min value of a `short`
`UCHAR MAX`	(≥255U)	max value of an `unsigned char`
`UINT_MAX`	(≥65535U)	max value of an `unsigned int`
`ULONG_MAX`	(≥4294967295U)	max value of an `unsigned long`
`USHRT_MAX`	(≥65535U)	max value of an `unsigned short`
Note: if the implementation treats `chars` as signed, then the values of `CHAR_MAX` and `CHAR_MIN` are the same as the equivalent `SCHAR` versions. If not, then the value of `CHAR_MIN` is zero and the value of `CHAR_MAX` is equal to the value of `UCHAR_MAX`.

Table 9.1. <limits.h>

9.5.2. Float.h

For floating point numbers, the file <float.h> contains a similar set of minimum values. (It is assumed that where no minimum value is specified, there is either no minimum, or the value depends on another value.)

Name	Allowable value	Comment
`FLT_RADIX`	(≥2)	the radix of exponent representation
`DBL_DIG`	(≥10)	the number of digits of precision in a `double`
`DBL_EPSILON`	(≤1E−9)	minimum positive number such that 1.0 + `x` ≠ 1.0
`DBL_MANT_DIG`	(—)	the number of base `FLT_RADIX` digits in the mantissa part of a `double`
`DBL_MAX`	(≥1E+37)	max value of a `double`
`DBL_MAX_10_EXP`	(≥+37)	max value of exponent (base 10) of a `double`
`DBL_MAX_EXP`	(—)	max value of exponent (base `FLT_RADIX`)) of a `double`
`DBL_MIN`	(≤1E−37)	min value of a `double`
`DBL_MIN_10_EXP`	(≤37)	minimum value of exponent (base 10) of a `double`
`DBL_MIN_EXP`	(—)	min value of exponent part of a `double` (base `FLT_RADIX`)
`FLT_DIG`	(≥6)	the number of digits of precision in a `float`
`FLT_EPSILON`	(≤1E−5)	minimum positive number such that 1.0 + x ≠ 1.0
`FLT_MANT_DIG`	(—)	the number of base `FLT_RADIX` digits in the mantissa of a `float`
`FLT_MAX`	(≥1E+37)	max value of a `float`
`FLT_MAX_10_EXP`	(≥+37)	max value (base 10) of exponent part of a `float`
`FLT_MAX_EXP`	(—)	max value (base `FLT_RADIX`) of exponent part of a `float`
`FLT_MIN`	(≤1E−37)	min value of a `float`
`FLT_MIN_10_EXP`	(≤−37)	min value (base 10) of exponent part of a `float`
`FLT_MIN_EXP`	(—)	min value (base `FLT_RADIX`) of exponent part of a `float`
`FLT_ROUNDS`	(0)	affects rounding of floating point addition: −1 indeterminate 0 towards zero 1 to nearest 2 towards +infinity 3 towards -infinity any other value is implementation defined.
`LDBL_DIG`	(≥10)	the number of digits of precision in a `long double`
`LDBL_EPSILON`	(≤1E−9)	minimum positive number such that 1.0 + x ≠= 1.0
`LDBL_MANT_DIG`	(—)	the number of base `FLT_RADIX` digits in the mantissa part of a `long double`
`LDBL_MAX`	(≥1E+37)	max value of a `long double`
`LDBL_MAX_10_EXP`	(≥+37)	max value of exponent (base 10) of a `long double`
`LDBL_MAX_EXP`	(—)	max value of exponent (base `FLT_RADIX`) of a `long double`
`LDBL_MIN`	(≤1E−37)	minimum value of a `long double`
`LDBL_MIN_10_EXP`	(≤−37)	min value of exponent part (base 10) of a `long double`
`LDBL_MIN_EXP`	(—)	min value of exponent part of a `long double` (base `FLT_RADIX`)

Table 9.2. <float.h>

9.5. Limits

9.5.1. Limits.h

9.5.2. Float.h

West Yorkshire Office

South East Regional Office