Movatterモバイル変換

[0]ホーム
URL:

cppreference.com
Namespaces
Variants
    Actions

      Floating-point extensions part 1: binary floating-point arithmetic

      From cppreference.com
      <c‎ |experimental
       
       
       
      Floating-point extensions part 1: Binary floating-point
       
      Merged into ISO C The functionality described on this page was merged into the mainline ISO C standard as of 3/2019; link TBD(since C23)

      Floating-point extensions for C - Part 1: Binary floating-point arithmetic, ISO/IEC TS 18661-1:2014, defines the following new components for the C standard library, as recommended by ISO/IEC/IEEE 60559:2011 (the current revision of IEEE-754)

      __STDC_IEC_60559_BFP__
      		 integer constant of typelong and value201ymmL, replaces__STDC_IEC_559__
      (macro constant)
      __STDC_IEC_60559_COMPLEX__
      		 integer constant of typelong and value201ymmL, replaces__STDC_IEC_559_COMPLEX__
      (macro constant)
      Defined in header<limits.h>
      CHAR_WIDTH SCHAR_WIDTH UCHAR_WIDTHSHRT_WIDTH USHRT_WIDTHINT_WIDTH UINT_WIDTHLONG_WIDTH ULONG_WIDTHLLONG_WIDTH ULLONG_WIDTH
      (FP Ext 1 TS)
      		 width, in bits, of the corresponding type
      (macro constant)
      Defined in header<float.h>
      (FP Ext 1 TS)
      		 conversions between all supported binary floating-point types and character sequences with at most CR_DECIMAL_DIG significant decimal digits are correctly rounded (this is at least DECIMAL_DIG + 3)
      (macro constant)
      Defined in header<fenv.h>
      femode_t
      (FP Ext 1 TS)
      		 collection of dynamic floating-point control modes supported by the implementation, including the dynamic rounding direction mode
      (typedef)
      FE_DFL_MODE
      (FP Ext 1 TS)
      		 pointer to the default femode_t
      (macro constant)
      FE_SNANS_ALWAYS_SIGNAL
      (FP Ext 1 TS)
      		 defined (as integer constant 1) if sNaN arguments cause the functions that suppress qNaNs, likehypot orfmax, to raise FE_INVALID and return a qNaN
      (macro constant)
      (FP Ext 1 TS)
      		 sets the specified floating-point exception flags without causing any side-effects that raising them would
      (function)
      (FP Ext 1 TS)
      		 tests if given flags are in a saved representation of the floating-point exception flags
      (function)
      (FP Ext 1 TS)
      		 gets and sets all the implementation’s dynamic floating-point control modes collectively
      (function)
      Defined in header<stdint.h>
      INTn_WIDTH UINTn_WIDTHINT_LEASTn_WIDTH UINT_LEASTn_WIDTHINT_FASTn_WIDTH UINT_FASTn_WIDTHINTPTR_WIDTH UINTPTR_WIDTHINTMAX_WIDTH UINTMAX_WIDTHPTRDIFF_WIDTHSIG_ATOMIC_WIDTHSIZE_WIDTHWCHAR_WIDTH WINT_WIDTH
      (FP Ext 1 TS)
      		 width, in bits, of the corresponding type
      (macro constant)
      Defined in header<stdlib.h>
      convert a single foating-point number to string using the specified snprintf format
      (function)
      Defined in header<math.h>
      FP_INT_UPWARDFP_INT_DOWNWARDFP_INT_TOWARDZEROFP_INT_TONEARESTFROMZEROFP_INT_TONEAREST
      (FP Ext 1 TS)
      		 rounding direction for the functions ceil, floor, trunc, round, and roundeven, suitable for use with fromfp family of functions
      (macro constant)
      FP_LLOGB0
      (FP Ext 1 TS)
      		 value returned byllogb if the argument is zero
      (macro constant)
      FP_LLOGBNAN
      (FP Ext 1 TS)
      		 value returned byllogb if the argument is NaN
      (macro constant)
      (FP Ext 1 TS)
      		 represents a signalling NaN for float, double, long double respectively
      (macro constant)
      FP_FAST_FADD FP_FAST_FADDL FP_FAST_DADDLFP_FAST_FSUB FP_FAST_FSUBL FP_FAST_DSUBLFP_FAST_FMUL FP_FAST_FMULL FP_FAST_DMULLFP_FAST_FDIV FP_FAST_FDIVL FP_FAST_DDIVLFP_FAST_FFMA FP_FAST_FFMAL FP_FAST_DFMALFP_FAST_FSQRT FP_FAST_FSQRTL FP_FAST_DSQRTL
      (FP Ext 1 TS)
      		 if defined, indicates that the corresponding function executes faster than the equivalent function in a larger type followed by a cast to target type
      (macro constant)
      iseqsig
      (FP Ext 1 TS)

      (function macro)
      iscanonical
      (FP Ext 1 TS)
      		 tests if the floating-point value is canonical
      (function macro)
      issignaling
      (FP Ext 1 TS)
      		 tests if the floating-point value is a signalling NaN
      (function macro)
      issubnormal
      (FP Ext 1 TS)
      		 tests if the floating-point value is subnormal
      (function macro)
      iszero
      (FP Ext 1 TS)
      		 tests if the floating-point value is a zero (positive, negative, unsigned)
      (function macro)
      round to signed integer using the specified rounding direction
      (function)
      round to unsigned integer using the specified rounding direction
      (function)
      round to signed integer using the specified rounding direction, reporting inexactness
      (function)
      round to unsigned integer using the specified rounding direction, reporting inexactness
      (function)
      rounds to nearest, halfway cases to even
      (function)
      (FP Ext 1 TS)
      		 equivalent tologb except the return type islong
      (function)
      returns the value of their argument of maximum magnitude
      (function)
      returns the value of their argument of minimum magnitude
      (function)
      returns the next greater representable floating-point value
      (function)
      returns the next smaller representable floating-point value
      (function)
      (FP Ext 1 TS)
      		 calculates x+y as if in infinite precision and rounded once to target type
      (function)
      (FP Ext 1 TS)
      		 calculates x-y as if in infinite precision and rounded once to target type
      (function)
      (FP Ext 1 TS)
      		 calculates x*y as if in infinite precision and rounded once to target type
      (function)
      (FP Ext 1 TS)
      		 calculates x/y as if in infinite precision and rounded once to target type
      (function)
      (FP Ext 1 TS)
      		 calculates the same asfma as if in infinite precision and rounded once to target type
      (function)
      (FP Ext 1 TS)
      		 calculates the same assqrt as if in infinite precision and rounded once to target type
      (function)
      orders two floating-point values using the ISO 60559 total order relation
      (function)
      orders the magnitudes of two floating-point values using the ISO 60559 total order relation
      (function)
      obtains the ISO 60559 canonical binary encoding of the given floating-point value
      (function)
      extracts the payload from the given NaN value
      (function)
      creates a quiet NaN with the specified payload
      (function)
      creates a signalling NaN with the specified payload
      (function)
      Defined in header<tgmath.h>
      roundeven
      (FP Ext 1 TS)
      		 generic overload ofroundeven
      (function)
      llogb
      (FP Ext 1 TS)
      		 generic overload ofllogb
      (function)
      fmaxmag
      (FP Ext 1 TS)
      		 generic overload offmaxmag
      (function)
      fminmag
      (FP Ext 1 TS)
      		 generic overload offminmag
      (function)
      nextup
      (FP Ext 1 TS)
      		 generic overload ofnextup
      (function)
      nextdown
      (FP Ext 1 TS)
      		 generic overload ofnextdown
      (function)
      fromfp
      (FP Ext 1 TS)
      		 generic overload offromfp
      (function)
      ufromfp
      (FP Ext 1 TS)
      		 generic overload ofufromfp
      (function)
      fromfpx
      (FP Ext 1 TS)
      		 generic overload offromfpx
      (function)
      ufromfpx
      (FP Ext 1 TS)
      		 generic overload ofufromfpx
      (function)
      nextdown
      (FP Ext 1 TS)
      		 generic overload ofnextdown
      (function)
      totalorder
      (FP Ext 1 TS)
      		 generic overload oftotalorder
      (function)
      totalordermag
      (FP Ext 1 TS)
      		 generic overload oftotalordermag
      (function)
      fadd
      (FP Ext 1 TS)
      		 generic overload offadd
      (function)
      dadd
      (FP Ext 1 TS)
      		 generic overload ofdadd
      (function)
      fsub
      (FP Ext 1 TS)
      		 generic overload offsub
      (function)
      dsub
      (FP Ext 1 TS)
      		 generic overload ofdsub
      (function)
      fmul
      (FP Ext 1 TS)
      		 generic overload offmul
      (function)
      dmul
      (FP Ext 1 TS)
      		 generic overload ofdmul
      (function)
      fdiv
      (FP Ext 1 TS)
      		 generic overload offdiv
      (function)
      ddiv
      (FP Ext 1 TS)
      		 generic overload ofddiv
      (function)
      ffma
      (FP Ext 1 TS)
      		 generic overload offfma
      (function)
      dfma
      (FP Ext 1 TS)
      		 generic overload ofdfma
      (function)
      fsqrt
      (FP Ext 1 TS)
      		 generic overload offsqrt
      (function)
      dsqrt
      (FP Ext 1 TS)
      		 generic overload ofdsqrt
      (function)

      [edit]Notes

      The standard C macros__STDC_IEC_559__ and__STDC_IEC_559_COMPLEX__ are made obsolete by this technical specification.

      All functions and macros added to the C library by this extension are only declared if a macro__STDC_WANT_IEC_60559_BFP_EXT__ is defined before the corresponding header is included.

      Besides additions to the standard library, ISO/IEC TS 18661-1:2014 makes a number of changes to the core language, notably splitting floating-point control between static (controlled by the new#pragma STDC FENV_ROUND), and dynamic (controlled byfesetround). Most math.h functions respect the static rounding mode, if set, over the dynamic rounding mode.

      This section is incomplete
      Reason: add to the pragma page or describe the pragma in full here?
      Retrieved from "https://en.cppreference.com/mwiki/index.php?title=c/experimental/fpext1&oldid=124001"
      [8]ページ先頭
      ©2009-2025 Movatter.jp