FLOAT - characteristics of floating point values.
Usage:
#include <float.h>
Description:
<float.h> defines manifests which describe how a
particular machine does floating point arithmetic. The model of a
floating point number X uses the following values:
- SIGN
- is +1 or -1.
- BASE
- is the base of the exponential representation of X. This
is an integer greater than 1.
- EXP
- is the exponent of X.
- EMAX
- is the maximum value of an exponent.
- EMIN
- is the minimum value of an exponent.
- PREC
- is the precision of X: the number of BASE digits in the
mantissa.
- DIG1,DIG2,...
- are the digits of the mantissa. All of these digits are
base BASE.
The value X is equal to the SIGN, times the BASE to the EXP,
times the mantissa.
Below we list the manifests defined in <float.h>.
- DBL_DIG
- the number of decimal digits of precision in a
"double" value.
- DBL_EPSILON
- the smallest positive "double" value that can
be added to 1.0 to get a value different from 1.0.
- DBL_MANT_DIG
- the number of base BASE digits in the mantissa of a
"double" value (i.e., the PREC).
- DBL_MAX
- the largest "double" value.
- DBL_MAX_10_EXP
- the largest integer such that 10 raised to that power is
representable as a "double" value.
- DBL_MAX_EXP
- the largest integer such that BASE raised to that power
minus 1 is representable as a "double" value.
- DBL_MIN
- the smallest normalized positive "double"
value.
- DBL_MIN_10_EXP
- the smallest negative integer such that 10 raised to the
power is still in the range of normalized
"double" values.
- DBL_MIN_EXP
- the smallest negative exponent for a "double"
value (i.e., EMIN).
- FLT_DIG
- the number of decimal digits of precision in a
"float" value.
- FLT_EPSILON
- the smallest positive "float" value that can be
added to 1.0F to get a number different from 1.0F.
- FLT_MANT_DIG
- the number of base BASE digits in the mantissa of a
"float" value (i.e., the PREC).
- FLT_MAX
- the largest "float" value.
- FLT_MAX_10_EXP
- the largest integer such that 10 raised to that power is
representable as a "float" value.
- FLT_MAX_EXP
- the largest integer such that BASE raised to that power
minus 1 is representable as a "float" value.
- FLT_MIN
- the smallest normalized positive "float" value.
- FLT_MIN_10_EXP
- the smallest negative integer such that 10 raised to the
power is still in the range of normalized
"float" values.
- FLT_MIN_EXP
- the smallest negative exponent for a "float"
value (i.e., EMIN).
- FLT_RADIX
- the BASE of a "float" number.
- FLT_ROUNDS
- indicates whether addition rounds or truncates. A
positive number indicates it rounds; 0 indicates it
truncates; -1 indicates the case is indeterminable.
- LDBL_DIG
- the number of decimal digits of precision in a "long
double" value.
- LDBL_EPSILON
- the smallest positive "long double" value that
can be added to 1.0L to get a value different from 1.0L.
- LDBL_MANT_DIG
- the number of base BASE digits in the mantissa of a
"long double" value (i.e., the PREC).
- LDBL_MAX
- the largest "long double" value.
- LDBL_MAX_10_EXP
- the largest integer such that 10 raised to that power is
representable as a "long double" value.
- LDBL_MAX_EXP
- the largest integer such that BASE raised to that power
minus 1 is representable as a "long double"
value.
- LDBL_MIN
- the smallest normalized positive "long double"
value.
- LDBL_MIN_10_EXP
- the smallest negative integer such that 10 raised to the
power is still in the range of normalized "long
double" values.
- LDBL_MIN_EXP
- the smallest negative exponent for a "long
double" value (i.e., EMIN).
Copyright © 1996, Thinkage Ltd.