newlib-cygwin/newlib/libm/mathfp/s_frexp.c

101 lines
2.1 KiB
C
Raw Normal View History

2000-02-18 03:39:52 +08:00
/* @(#)z_frexp.c 1.0 98/08/13 */
/*
FUNCTION
<<frexp>>, <<frexpf>>---split floating-point number
INDEX
frexp
INDEX
frexpf
SYNOPSIS
2000-02-18 03:39:52 +08:00
#include <math.h>
double frexp(double <[val]>, int *<[exp]>);
float frexpf(float <[val]>, int *<[exp]>);
DESCRIPTION
2003-10-20 Bob Wilson <bob.wilson@acm.org> * libc/locale/locale.c: Use double quotes in code. * libc/locale/locale.tex: Likewise. * libc/libc.texinfo: Hyphenate "floating-point". * libc/machine/necv70/necv70.tex: Likewise. * libc/stdio/sprintf.c: Likewise. * libc/stdio/sscanf.c: Likewise. * libc/stdlib/atof.c: Likewise. * libc/stdlib/ldtoa.c: Fix comment typo. * libc/stdlib/div.c: Use @ifnottex instead of @ifinfo. * libc/stdlib/ldiv.c: Likewise. * libm/common/s_expm1.c: Likewise. * libm/common/s_ilogb.c: Hyphenate "floating-point". Use "nonzero". * libm/common/s_infinity: Hyphenate "double- and single-precision". * libm/common/s_nan.c: Likewise. Also correct the FUNCTION summary. * libm/common/s_nextafter.c: Fix typo. Hyphenate "double-precision" and "floating-point". * libm/common/s_scalbn.c: Correct the FUNCTION summary. * libm/math/e_pow.c: Fix comment typo. * libm/math/math.tex: Hyphenate "single-precision" and "floating-point". * libm/math/s_asinh.c: Use @ifnottex instead of @ifinfo. * libm/math/s_atan.c: Likewise. * libm/math/s_frexp.c: Likewise. Also use "nonzero". * libm/math/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/math/w_acos.c: Likewise. * libm/math/w_acosh.c: Likewise. * libm/math/w_asin.c: Likewise. * libm/math/w_atan2.c: Likewise. * libm/math/w_atanh.c: Likewise. * libm/math/w_cosh.c: Likewise. * libm/math/w_exp.c: Likewise. * libm/math/w_exp2.c: Likewise. Add missing @end and missing @tex version of 2^x in the description. * libm/math/w_fmod.c: Use @ifnottex instead of @ifinfo. * libm/math/w_gamma.c: Likewise. * libm/math/w_hypot.c: Likewise. * libm/math/w_j0.c: Likewise. * libm/math/w_sinh.c: Likewise. * libm/math/w_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/mathfp.tex: Hyphenate "single-precision" and "floating-point". * libm/mathfp/e_acosh.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/e_atanh.c: Likewise. * libm/mathfp/e_hypot.c: Likewise. * libm/mathfp/er_lgamma.c: Likewise. * libm/mathfp/s_acos.c: Likewise. * libm/mathfp/s_asine.c: Likewise. * libm/mathfp/s_asinh.c: Likewise. * libm/mathfp/s_atan.c: Likewise. * libm/mathfp/s_atan2.c: Likewise. * libm/mathfp/s_atangent.c: Likewise. * libm/mathfp/s_cosh.c: Likewise. * libm/mathfp/s_exp.c: Likewise. * libm/mathfp/s_fmod.c: Likewise. * libm/mathfp/s_frexp.c: Likewise. Also use "nonzero". * libm/mathfp/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/s_sineh.c: Likewise. * libm/mathfp/w_jn.c: Likewise.
2003-10-21 02:46:38 +08:00
All nonzero, normal numbers can be described as <[m]> * 2**<[p]>.
2000-02-18 03:39:52 +08:00
<<frexp>> represents the double <[val]> as a mantissa <[m]>
and a power of two <[p]>. The resulting mantissa will always
be greater than or equal to <<0.5>>, and less than <<1.0>> (as
long as <[val]> is nonzero). The power of two will be stored
in <<*>><[exp]>.
2003-10-20 Bob Wilson <bob.wilson@acm.org> * libc/locale/locale.c: Use double quotes in code. * libc/locale/locale.tex: Likewise. * libc/libc.texinfo: Hyphenate "floating-point". * libc/machine/necv70/necv70.tex: Likewise. * libc/stdio/sprintf.c: Likewise. * libc/stdio/sscanf.c: Likewise. * libc/stdlib/atof.c: Likewise. * libc/stdlib/ldtoa.c: Fix comment typo. * libc/stdlib/div.c: Use @ifnottex instead of @ifinfo. * libc/stdlib/ldiv.c: Likewise. * libm/common/s_expm1.c: Likewise. * libm/common/s_ilogb.c: Hyphenate "floating-point". Use "nonzero". * libm/common/s_infinity: Hyphenate "double- and single-precision". * libm/common/s_nan.c: Likewise. Also correct the FUNCTION summary. * libm/common/s_nextafter.c: Fix typo. Hyphenate "double-precision" and "floating-point". * libm/common/s_scalbn.c: Correct the FUNCTION summary. * libm/math/e_pow.c: Fix comment typo. * libm/math/math.tex: Hyphenate "single-precision" and "floating-point". * libm/math/s_asinh.c: Use @ifnottex instead of @ifinfo. * libm/math/s_atan.c: Likewise. * libm/math/s_frexp.c: Likewise. Also use "nonzero". * libm/math/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/math/w_acos.c: Likewise. * libm/math/w_acosh.c: Likewise. * libm/math/w_asin.c: Likewise. * libm/math/w_atan2.c: Likewise. * libm/math/w_atanh.c: Likewise. * libm/math/w_cosh.c: Likewise. * libm/math/w_exp.c: Likewise. * libm/math/w_exp2.c: Likewise. Add missing @end and missing @tex version of 2^x in the description. * libm/math/w_fmod.c: Use @ifnottex instead of @ifinfo. * libm/math/w_gamma.c: Likewise. * libm/math/w_hypot.c: Likewise. * libm/math/w_j0.c: Likewise. * libm/math/w_sinh.c: Likewise. * libm/math/w_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/mathfp.tex: Hyphenate "single-precision" and "floating-point". * libm/mathfp/e_acosh.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/e_atanh.c: Likewise. * libm/mathfp/e_hypot.c: Likewise. * libm/mathfp/er_lgamma.c: Likewise. * libm/mathfp/s_acos.c: Likewise. * libm/mathfp/s_asine.c: Likewise. * libm/mathfp/s_asinh.c: Likewise. * libm/mathfp/s_atan.c: Likewise. * libm/mathfp/s_atan2.c: Likewise. * libm/mathfp/s_atangent.c: Likewise. * libm/mathfp/s_cosh.c: Likewise. * libm/mathfp/s_exp.c: Likewise. * libm/mathfp/s_fmod.c: Likewise. * libm/mathfp/s_frexp.c: Likewise. Also use "nonzero". * libm/mathfp/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/s_sineh.c: Likewise. * libm/mathfp/w_jn.c: Likewise.
2003-10-21 02:46:38 +08:00
@ifnottex
2000-02-18 03:39:52 +08:00
<[m]> and <[p]> are calculated so that
<[val]> is <[m]> times <<2>> to the power <[p]>.
2003-10-20 Bob Wilson <bob.wilson@acm.org> * libc/locale/locale.c: Use double quotes in code. * libc/locale/locale.tex: Likewise. * libc/libc.texinfo: Hyphenate "floating-point". * libc/machine/necv70/necv70.tex: Likewise. * libc/stdio/sprintf.c: Likewise. * libc/stdio/sscanf.c: Likewise. * libc/stdlib/atof.c: Likewise. * libc/stdlib/ldtoa.c: Fix comment typo. * libc/stdlib/div.c: Use @ifnottex instead of @ifinfo. * libc/stdlib/ldiv.c: Likewise. * libm/common/s_expm1.c: Likewise. * libm/common/s_ilogb.c: Hyphenate "floating-point". Use "nonzero". * libm/common/s_infinity: Hyphenate "double- and single-precision". * libm/common/s_nan.c: Likewise. Also correct the FUNCTION summary. * libm/common/s_nextafter.c: Fix typo. Hyphenate "double-precision" and "floating-point". * libm/common/s_scalbn.c: Correct the FUNCTION summary. * libm/math/e_pow.c: Fix comment typo. * libm/math/math.tex: Hyphenate "single-precision" and "floating-point". * libm/math/s_asinh.c: Use @ifnottex instead of @ifinfo. * libm/math/s_atan.c: Likewise. * libm/math/s_frexp.c: Likewise. Also use "nonzero". * libm/math/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/math/w_acos.c: Likewise. * libm/math/w_acosh.c: Likewise. * libm/math/w_asin.c: Likewise. * libm/math/w_atan2.c: Likewise. * libm/math/w_atanh.c: Likewise. * libm/math/w_cosh.c: Likewise. * libm/math/w_exp.c: Likewise. * libm/math/w_exp2.c: Likewise. Add missing @end and missing @tex version of 2^x in the description. * libm/math/w_fmod.c: Use @ifnottex instead of @ifinfo. * libm/math/w_gamma.c: Likewise. * libm/math/w_hypot.c: Likewise. * libm/math/w_j0.c: Likewise. * libm/math/w_sinh.c: Likewise. * libm/math/w_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/mathfp.tex: Hyphenate "single-precision" and "floating-point". * libm/mathfp/e_acosh.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/e_atanh.c: Likewise. * libm/mathfp/e_hypot.c: Likewise. * libm/mathfp/er_lgamma.c: Likewise. * libm/mathfp/s_acos.c: Likewise. * libm/mathfp/s_asine.c: Likewise. * libm/mathfp/s_asinh.c: Likewise. * libm/mathfp/s_atan.c: Likewise. * libm/mathfp/s_atan2.c: Likewise. * libm/mathfp/s_atangent.c: Likewise. * libm/mathfp/s_cosh.c: Likewise. * libm/mathfp/s_exp.c: Likewise. * libm/mathfp/s_fmod.c: Likewise. * libm/mathfp/s_frexp.c: Likewise. Also use "nonzero". * libm/mathfp/s_isnan.c: Hyphenate "single-precision" and "floating-point". Use "nonzero". * libm/math/s_pow.c: Replace "exp1.0nt" with "exponent". * libm/mathfp/s_ldexp.c: Use @ifnottex instead of @ifinfo. * libm/mathfp/s_sineh.c: Likewise. * libm/mathfp/w_jn.c: Likewise.
2003-10-21 02:46:38 +08:00
@end ifnottex
2000-02-18 03:39:52 +08:00
@tex
<[m]> and <[p]> are calculated so that
$ val = m \times 2^p $.
@end tex
<<frexpf>> is identical, other than taking and returning
floats rather than doubles.
RETURNS
<<frexp>> returns the mantissa <[m]>. If <[val]> is <<0>>, infinity,
or Nan, <<frexp>> will set <<*>><[exp]> to <<0>> and return <[val]>.
PORTABILITY
<<frexp>> is ANSI.
<<frexpf>> is an extension.
*/
/*****************************************************************
* frexp
*
* Input:
* d - floating point value
* exp - exponent value
*
* Output:
* A floating point value in the range [0.5, 1).
*
* Description:
* This routine breaks a floating point value into a number f and
* an exponent exp such that d = f * 2 ^ exp.
*
*****************************************************************/
#include "fdlibm.h"
#include "zmath.h"
#ifndef _DOUBLE_IS_32BITS
double frexp (double d, int *exp)
{
double f;
__uint32_t hd, ld, hf, lf;
/* Check for special values. */
switch (numtest (d))
{
case NAN:
case INF:
errno = EDOM;
case 0:
*exp = 0;
return (d);
}
2000-02-18 03:39:52 +08:00
EXTRACT_WORDS (hd, ld, d);
/* Get the exponent. */
*exp = ((hd & 0x7ff00000) >> 20) - 1022;
/* Get the mantissa. */
lf = ld;
hf = hd & 0x800fffff;
hf |= 0x3fe00000;
INSERT_WORDS (f, hf, lf);
return (f);
}
#endif /* _DOUBLE_IS_32BITS */