gl/float+.h

   1 /* Supplemental information about the floating-point formats.
   2    Copyright (C) 2007, 2009-2024 Free Software Foundation, Inc.
   3    Written by Bruno Haible <bruno@clisp.org>, 2007.
   4
   5    This file is free software: you can redistribute it and/or modify
   6    it under the terms of the GNU Lesser General Public License as
   7    published by the Free Software Foundation; either version 2.1 of the
   8    License, or (at your option) any later version.
   9
  10    This file is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13    GNU Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public License
  16    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  17
  18 #ifndef _FLOATPLUS_H
  19 #define _FLOATPLUS_H
  20
  21 #include <float.h>
  22 #include <limits.h>
  23
  24 /* Number of bits in the mantissa of a floating-point number, including the
  25    "hidden bit".  */
  26 #if FLT_RADIX == 2
  27 # define FLT_MANT_BIT FLT_MANT_DIG
  28 # define DBL_MANT_BIT DBL_MANT_DIG
  29 # define LDBL_MANT_BIT LDBL_MANT_DIG
  30 #elif FLT_RADIX == 4
  31 # define FLT_MANT_BIT (FLT_MANT_DIG * 2)
  32 # define DBL_MANT_BIT (DBL_MANT_DIG * 2)
  33 # define LDBL_MANT_BIT (LDBL_MANT_DIG * 2)
  34 #elif FLT_RADIX == 16
  35 # define FLT_MANT_BIT (FLT_MANT_DIG * 4)
  36 # define DBL_MANT_BIT (DBL_MANT_DIG * 4)
  37 # define LDBL_MANT_BIT (LDBL_MANT_DIG * 4)
  38 #endif
  39
  40 /* Bit mask that can be used to mask the exponent, as an unsigned number.  */
  41 #define FLT_EXP_MASK ((FLT_MAX_EXP - FLT_MIN_EXP) | 7)
  42 #define DBL_EXP_MASK ((DBL_MAX_EXP - DBL_MIN_EXP) | 7)
  43 #define LDBL_EXP_MASK ((LDBL_MAX_EXP - LDBL_MIN_EXP) | 7)
  44
  45 /* Number of bits used for the exponent of a floating-point number, including
  46    the exponent's sign.  */
  47 #define FLT_EXP_BIT \
  48   (FLT_EXP_MASK < 0x100 ? 8 : \
  49    FLT_EXP_MASK < 0x200 ? 9 : \
  50    FLT_EXP_MASK < 0x400 ? 10 : \
  51    FLT_EXP_MASK < 0x800 ? 11 : \
  52    FLT_EXP_MASK < 0x1000 ? 12 : \
  53    FLT_EXP_MASK < 0x2000 ? 13 : \
  54    FLT_EXP_MASK < 0x4000 ? 14 : \
  55    FLT_EXP_MASK < 0x8000 ? 15 : \
  56    FLT_EXP_MASK < 0x10000 ? 16 : \
  57    FLT_EXP_MASK < 0x20000 ? 17 : \
  58    FLT_EXP_MASK < 0x40000 ? 18 : \
  59    FLT_EXP_MASK < 0x80000 ? 19 : \
  60    FLT_EXP_MASK < 0x100000 ? 20 : \
  61    FLT_EXP_MASK < 0x200000 ? 21 : \
  62    FLT_EXP_MASK < 0x400000 ? 22 : \
  63    FLT_EXP_MASK < 0x800000 ? 23 : \
  64    FLT_EXP_MASK < 0x1000000 ? 24 : \
  65    FLT_EXP_MASK < 0x2000000 ? 25 : \
  66    FLT_EXP_MASK < 0x4000000 ? 26 : \
  67    FLT_EXP_MASK < 0x8000000 ? 27 : \
  68    FLT_EXP_MASK < 0x10000000 ? 28 : \
  69    FLT_EXP_MASK < 0x20000000 ? 29 : \
  70    FLT_EXP_MASK < 0x40000000 ? 30 : \
  71    FLT_EXP_MASK <= 0x7fffffff ? 31 : \
  72    32)
  73 #define DBL_EXP_BIT \
  74   (DBL_EXP_MASK < 0x100 ? 8 : \
  75    DBL_EXP_MASK < 0x200 ? 9 : \
  76    DBL_EXP_MASK < 0x400 ? 10 : \
  77    DBL_EXP_MASK < 0x800 ? 11 : \
  78    DBL_EXP_MASK < 0x1000 ? 12 : \
  79    DBL_EXP_MASK < 0x2000 ? 13 : \
  80    DBL_EXP_MASK < 0x4000 ? 14 : \
  81    DBL_EXP_MASK < 0x8000 ? 15 : \
  82    DBL_EXP_MASK < 0x10000 ? 16 : \
  83    DBL_EXP_MASK < 0x20000 ? 17 : \
  84    DBL_EXP_MASK < 0x40000 ? 18 : \
  85    DBL_EXP_MASK < 0x80000 ? 19 : \
  86    DBL_EXP_MASK < 0x100000 ? 20 : \
  87    DBL_EXP_MASK < 0x200000 ? 21 : \
  88    DBL_EXP_MASK < 0x400000 ? 22 : \
  89    DBL_EXP_MASK < 0x800000 ? 23 : \
  90    DBL_EXP_MASK < 0x1000000 ? 24 : \
  91    DBL_EXP_MASK < 0x2000000 ? 25 : \
  92    DBL_EXP_MASK < 0x4000000 ? 26 : \
  93    DBL_EXP_MASK < 0x8000000 ? 27 : \
  94    DBL_EXP_MASK < 0x10000000 ? 28 : \
  95    DBL_EXP_MASK < 0x20000000 ? 29 : \
  96    DBL_EXP_MASK < 0x40000000 ? 30 : \
  97    DBL_EXP_MASK <= 0x7fffffff ? 31 : \
  98    32)
  99 #define LDBL_EXP_BIT \
 100   (LDBL_EXP_MASK < 0x100 ? 8 : \
 101    LDBL_EXP_MASK < 0x200 ? 9 : \
 102    LDBL_EXP_MASK < 0x400 ? 10 : \
 103    LDBL_EXP_MASK < 0x800 ? 11 : \
 104    LDBL_EXP_MASK < 0x1000 ? 12 : \
 105    LDBL_EXP_MASK < 0x2000 ? 13 : \
 106    LDBL_EXP_MASK < 0x4000 ? 14 : \
 107    LDBL_EXP_MASK < 0x8000 ? 15 : \
 108    LDBL_EXP_MASK < 0x10000 ? 16 : \
 109    LDBL_EXP_MASK < 0x20000 ? 17 : \
 110    LDBL_EXP_MASK < 0x40000 ? 18 : \
 111    LDBL_EXP_MASK < 0x80000 ? 19 : \
 112    LDBL_EXP_MASK < 0x100000 ? 20 : \
 113    LDBL_EXP_MASK < 0x200000 ? 21 : \
 114    LDBL_EXP_MASK < 0x400000 ? 22 : \
 115    LDBL_EXP_MASK < 0x800000 ? 23 : \
 116    LDBL_EXP_MASK < 0x1000000 ? 24 : \
 117    LDBL_EXP_MASK < 0x2000000 ? 25 : \
 118    LDBL_EXP_MASK < 0x4000000 ? 26 : \
 119    LDBL_EXP_MASK < 0x8000000 ? 27 : \
 120    LDBL_EXP_MASK < 0x10000000 ? 28 : \
 121    LDBL_EXP_MASK < 0x20000000 ? 29 : \
 122    LDBL_EXP_MASK < 0x40000000 ? 30 : \
 123    LDBL_EXP_MASK <= 0x7fffffff ? 31 : \
 124    32)
 125
 126 /* Number of bits used for a floating-point number: the mantissa (not
 127    counting the "hidden bit", since it may or may not be explicit), the
 128    exponent, and the sign.  */
 129 #define FLT_TOTAL_BIT ((FLT_MANT_BIT - 1) + FLT_EXP_BIT + 1)
 130 #define DBL_TOTAL_BIT ((DBL_MANT_BIT - 1) + DBL_EXP_BIT + 1)
 131 #define LDBL_TOTAL_BIT ((LDBL_MANT_BIT - 1) + LDBL_EXP_BIT + 1)
 132
 133 /* Number of bytes used for a floating-point number.
 134    This can be smaller than the 'sizeof'.  For example, on i386 systems,
 135    'long double' most often have LDBL_MANT_BIT = 64, LDBL_EXP_BIT = 16, hence
 136    LDBL_TOTAL_BIT = 80 bits, i.e. 10 bytes of consecutive memory, but
 137    sizeof (long double) = 12 or = 16.  */
 138 #define SIZEOF_FLT ((FLT_TOTAL_BIT + CHAR_BIT - 1) / CHAR_BIT)
 139 #define SIZEOF_DBL ((DBL_TOTAL_BIT + CHAR_BIT - 1) / CHAR_BIT)
 140 #define SIZEOF_LDBL ((LDBL_TOTAL_BIT + CHAR_BIT - 1) / CHAR_BIT)
 141
 142 /* Verify that SIZEOF_FLT <= sizeof (float) etc.  */
 143 typedef int verify_sizeof_flt[SIZEOF_FLT <= sizeof (float) ? 1 : -1];
 144 typedef int verify_sizeof_dbl[SIZEOF_DBL <= sizeof (double) ? 1 : - 1];
 145 typedef int verify_sizeof_ldbl[SIZEOF_LDBL <= sizeof (long double) ? 1 : - 1];
 146
 147 #endif /* _FLOATPLUS_H */