float.h -- Header File

Define constants for floating-point numbers

The  header file  float.h defines the  following manifest  constants, which
mark the  limits for computation  of floating-point numbers.   The prefixes
DBL, FLT, and LDBL refer, respective, to double, float, and long double:

DBL_DIG
     Number of decimal digits of precision.

DBL_EPSILON
     Smallest possible  floating-point number x, such that  1.0 plus x does
     not test equal to 1.0.

DBL_MANT_DIG
     Number of digits in the floating-point mantissa for base FLT_RADIX.

DBL_MAX
     Largest number that can be held by type double.

DBL_MAX_EXP
     Largest integer  such that the value of FLT_RADIX  raised to its power
     minus one is less than or equal to DBL_MAX.

DBL_MAX_10_EXP
     Largest integer  such that  ten raised  to its power  is less  than or
     equal to DBL_MAX.

DBL_MIN
     Smallest number that can be held by type double.

DBL_MIN_EXP
     Smallest integer such that the  value of FLT_RADIX raised to its power
     minus one is greater than or equal to DBL_MIN.

DBL_MIN_10_EXP
     Smallest integer such that ten raised  to its power is greater than or
     equal to DBL_MAX.

FLT_DIG
     Number of decimal digits of precision.

FLT_EPSILON
     Smallest floating-point  number x, such that 1.0 plus  x does not test
     equal to 1.0.

FLT_MANT_DIG
     Number of digits in the floating-point mantissa for base FLT_RADIX.

FLT_MAX
     Largest number that can be held by type float.

FLT_MAX_EXP
     Largest integer  such that the value of FLT_RADIX  raised to its power
     minus one is less than or equal to FLT_MAX.

FLT_MAX_10_EXP
     Largest integer  such that  ten raised  to its power  is less  than or
     equal to FLT_MAX.

FLT_MIN
     Smallest number that can be held by type float.

FLT_MIN_EXP
     Smallest integer such that the  value of FLT_RADIX raised to its power
     minus one is greater than or equal to FLT_MIN.

FLT_MIN_10_EXP
     Smallest integer such that ten raised  to its power is greater than or
     equal to FLT_MIN.

FLT_RADIX
     Base  in  which  the  exponents  of  all  floating-point  numbers  are
     represented.

FLT_ROUNDS
     Manner  of rounding  used by  the  implementation.  The  ANSI Standard
     defines the rounding codes as follows:

    -1   Indeterminable, i.e., no strict rules apply
     0   Toward zero, i.e., truncation
     1   To nearest, i.e., rounds to nearest representable value
     2   Toward positive infinity, i.e., always rounds up
     3   Toward negative infinity, i.e., always rounds down

     COHERENT uses type-1 rounding.

LDBL_DIG
     Number of decimal digits of precision.

LDBL_EPSILON
     Smallest floating-point  number x, such that 1.0 plus  x does not test
     equal to 1.0.

LDBL_MANT_DIG
     Number of digits in the floating-point mantissa for base FLT_RADIX.

LDBL_MAX
     Largest number that can be held by type long double.

LDBL_MAX_EXP
     Largest integer  such that the value of FLT_RADIX  raised to its power
     minus one is less than or equal to LDBL_MAX.

LDBL_MAX_10_EXP
     Largest integer  such that  ten raised  to its power  is less  than or
     equal to LDBL_MAX.

LDBL_MIN
     Smallest  number that  can be  held by  type long  double. Must  be no
     greater than 1E-37.

LDBL_MIN_EXP
     Smallest integer such that the  value of FLT_RADIX raised to its power
     minus one is greater than or equal to LDBL_MIN.

LDBL_MIN_10_EXP
     Smallest integer such that ten raised  to its power is greater than or
     equal to LDBL_MIN.

See Also

double,
float,
header files
ANSI Standard, §5.2.4.2.2

Notes

COHERENT's C compiler does not yet implement type long double.

© 2012 by Stephen A. Ness