Home | History | Annotate | Download | only in templates
      1 
      2 /*
      3 ===============================================================================
      4 
      5 This C header file is part of the SoftFloat IEC/IEEE Floating-point
      6 Arithmetic Package, Release 2a.
      7 
      8 Written by John R. Hauser.  This work was made possible in part by the
      9 International Computer Science Institute, located at Suite 600, 1947 Center
     10 Street, Berkeley, California 94704.  Funding was partially provided by the
     11 National Science Foundation under grant MIP-9311980.  The original version
     12 of this code was written as part of a project to build a fixed-point vector
     13 processor in collaboration with the University of California at Berkeley,
     14 overseen by Profs. Nelson Morgan and John Wawrzynek.  More information
     15 is available through the Web page `http://HTTP.CS.Berkeley.EDU/~jhauser/
     16 arithmetic/SoftFloat.html'.
     17 
     18 THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
     19 has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
     20 TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
     21 PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
     22 AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
     23 
     24 Derivative works are acceptable, even for commercial purposes, so long as
     25 (1) they include prominent notice that the work is derivative, and (2) they
     26 include prominent notice akin to these four paragraphs for those parts of
     27 this code that are retained.
     28 
     29 ===============================================================================
     30 */
     31 
     32 /*
     33 -------------------------------------------------------------------------------
     34 The macro `FLOATX80' must be defined to enable the extended double-precision
     35 floating-point format `floatx80'.  If this macro is not defined, the
     36 `floatx80' type will not be defined, and none of the functions that either
     37 input or output the `floatx80' type will be defined.  The same applies to
     38 the `FLOAT128' macro and the quadruple-precision format `float128'.
     39 -------------------------------------------------------------------------------
     40 */
     41 #define FLOATX80
     42 #define FLOAT128
     43 
     44 /*
     45 -------------------------------------------------------------------------------
     46 Software IEC/IEEE floating-point types.
     47 -------------------------------------------------------------------------------
     48 */
     49 typedef !!!bits32 float32;
     50 typedef !!!bits64 float64;
     51 #ifdef FLOATX80
     52 typedef struct {
     53     !!!bits16 high;
     54     !!!bits64 low;
     55 } floatx80;
     56 #endif
     57 #ifdef FLOAT128
     58 typedef struct {
     59     !!!bits64 high, low;
     60 } float128;
     61 #endif
     62 
     63 /*
     64 -------------------------------------------------------------------------------
     65 Software IEC/IEEE floating-point underflow tininess-detection mode.
     66 -------------------------------------------------------------------------------
     67 */
     68 extern !!!int8 float_detect_tininess;
     69 enum {
     70     float_tininess_after_rounding  = 0,
     71     float_tininess_before_rounding = 1
     72 };
     73 
     74 /*
     75 -------------------------------------------------------------------------------
     76 Software IEC/IEEE floating-point rounding mode.
     77 -------------------------------------------------------------------------------
     78 */
     79 extern !!!int8 float_rounding_mode;
     80 enum {
     81     float_round_nearest_even = 0,
     82     float_round_to_zero      = 1,
     83     float_round_down         = 2,
     84     float_round_up           = 3
     85 };
     86 
     87 /*
     88 -------------------------------------------------------------------------------
     89 Software IEC/IEEE floating-point exception flags.
     90 -------------------------------------------------------------------------------
     91 */
     92 extern !!!int8 float_exception_flags;
     93 enum {
     94     float_flag_inexact   =  1,
     95     float_flag_underflow =  2,
     96     float_flag_overflow  =  4,
     97     float_flag_divbyzero =  8,
     98     float_flag_invalid   = 16
     99 };
    100 
    101 /*
    102 -------------------------------------------------------------------------------
    103 Routine to raise any or all of the software IEC/IEEE floating-point
    104 exception flags.
    105 -------------------------------------------------------------------------------
    106 */
    107 void float_raise( !!!int8 );
    108 
    109 /*
    110 -------------------------------------------------------------------------------
    111 Software IEC/IEEE integer-to-floating-point conversion routines.
    112 -------------------------------------------------------------------------------
    113 */
    114 float32 int32_to_float32( !!!int32 );
    115 float64 int32_to_float64( !!!int32 );
    116 #ifdef FLOATX80
    117 floatx80 int32_to_floatx80( !!!int32 );
    118 #endif
    119 #ifdef FLOAT128
    120 float128 int32_to_float128( !!!int32 );
    121 #endif
    122 float32 int64_to_float32( !!!int64 );
    123 float64 int64_to_float64( !!!int64 );
    124 #ifdef FLOATX80
    125 floatx80 int64_to_floatx80( !!!int64 );
    126 #endif
    127 #ifdef FLOAT128
    128 float128 int64_to_float128( !!!int64 );
    129 #endif
    130 
    131 /*
    132 -------------------------------------------------------------------------------
    133 Software IEC/IEEE single-precision conversion routines.
    134 -------------------------------------------------------------------------------
    135 */
    136 !!!int32 float32_to_int32( float32 );
    137 !!!int32 float32_to_int32_round_to_zero( float32 );
    138 !!!int64 float32_to_int64( float32 );
    139 !!!int64 float32_to_int64_round_to_zero( float32 );
    140 float64 float32_to_float64( float32 );
    141 #ifdef FLOATX80
    142 floatx80 float32_to_floatx80( float32 );
    143 #endif
    144 #ifdef FLOAT128
    145 float128 float32_to_float128( float32 );
    146 #endif
    147 
    148 /*
    149 -------------------------------------------------------------------------------
    150 Software IEC/IEEE single-precision operations.
    151 -------------------------------------------------------------------------------
    152 */
    153 float32 float32_round_to_int( float32 );
    154 float32 float32_add( float32, float32 );
    155 float32 float32_sub( float32, float32 );
    156 float32 float32_mul( float32, float32 );
    157 float32 float32_div( float32, float32 );
    158 float32 float32_rem( float32, float32 );
    159 float32 float32_sqrt( float32 );
    160 !!!flag float32_eq( float32, float32 );
    161 !!!flag float32_le( float32, float32 );
    162 !!!flag float32_lt( float32, float32 );
    163 !!!flag float32_eq_signaling( float32, float32 );
    164 !!!flag float32_le_quiet( float32, float32 );
    165 !!!flag float32_lt_quiet( float32, float32 );
    166 !!!flag float32_is_signaling_nan( float32 );
    167 
    168 /*
    169 -------------------------------------------------------------------------------
    170 Software IEC/IEEE double-precision conversion routines.
    171 -------------------------------------------------------------------------------
    172 */
    173 !!!int32 float64_to_int32( float64 );
    174 !!!int32 float64_to_int32_round_to_zero( float64 );
    175 !!!int64 float64_to_int64( float64 );
    176 !!!int64 float64_to_int64_round_to_zero( float64 );
    177 float32 float64_to_float32( float64 );
    178 #ifdef FLOATX80
    179 floatx80 float64_to_floatx80( float64 );
    180 #endif
    181 #ifdef FLOAT128
    182 float128 float64_to_float128( float64 );
    183 #endif
    184 
    185 /*
    186 -------------------------------------------------------------------------------
    187 Software IEC/IEEE double-precision operations.
    188 -------------------------------------------------------------------------------
    189 */
    190 float64 float64_round_to_int( float64 );
    191 float64 float64_add( float64, float64 );
    192 float64 float64_sub( float64, float64 );
    193 float64 float64_mul( float64, float64 );
    194 float64 float64_div( float64, float64 );
    195 float64 float64_rem( float64, float64 );
    196 float64 float64_sqrt( float64 );
    197 !!!flag float64_eq( float64, float64 );
    198 !!!flag float64_le( float64, float64 );
    199 !!!flag float64_lt( float64, float64 );
    200 !!!flag float64_eq_signaling( float64, float64 );
    201 !!!flag float64_le_quiet( float64, float64 );
    202 !!!flag float64_lt_quiet( float64, float64 );
    203 !!!flag float64_is_signaling_nan( float64 );
    204 
    205 #ifdef FLOATX80
    206 
    207 /*
    208 -------------------------------------------------------------------------------
    209 Software IEC/IEEE extended double-precision conversion routines.
    210 -------------------------------------------------------------------------------
    211 */
    212 !!!int32 floatx80_to_int32( floatx80 );
    213 !!!int32 floatx80_to_int32_round_to_zero( floatx80 );
    214 !!!int64 floatx80_to_int64( floatx80 );
    215 !!!int64 floatx80_to_int64_round_to_zero( floatx80 );
    216 float32 floatx80_to_float32( floatx80 );
    217 float64 floatx80_to_float64( floatx80 );
    218 #ifdef FLOAT128
    219 float128 floatx80_to_float128( floatx80 );
    220 #endif
    221 
    222 /*
    223 -------------------------------------------------------------------------------
    224 Software IEC/IEEE extended double-precision rounding precision.  Valid
    225 values are 32, 64, and 80.
    226 -------------------------------------------------------------------------------
    227 */
    228 extern !!!int8 floatx80_rounding_precision;
    229 
    230 /*
    231 -------------------------------------------------------------------------------
    232 Software IEC/IEEE extended double-precision operations.
    233 -------------------------------------------------------------------------------
    234 */
    235 floatx80 floatx80_round_to_int( floatx80 );
    236 floatx80 floatx80_add( floatx80, floatx80 );
    237 floatx80 floatx80_sub( floatx80, floatx80 );
    238 floatx80 floatx80_mul( floatx80, floatx80 );
    239 floatx80 floatx80_div( floatx80, floatx80 );
    240 floatx80 floatx80_rem( floatx80, floatx80 );
    241 floatx80 floatx80_sqrt( floatx80 );
    242 !!!flag floatx80_eq( floatx80, floatx80 );
    243 !!!flag floatx80_le( floatx80, floatx80 );
    244 !!!flag floatx80_lt( floatx80, floatx80 );
    245 !!!flag floatx80_eq_signaling( floatx80, floatx80 );
    246 !!!flag floatx80_le_quiet( floatx80, floatx80 );
    247 !!!flag floatx80_lt_quiet( floatx80, floatx80 );
    248 !!!flag floatx80_is_signaling_nan( floatx80 );
    249 
    250 #endif
    251 
    252 #ifdef FLOAT128
    253 
    254 /*
    255 -------------------------------------------------------------------------------
    256 Software IEC/IEEE quadruple-precision conversion routines.
    257 -------------------------------------------------------------------------------
    258 */
    259 !!!int32 float128_to_int32( float128 );
    260 !!!int32 float128_to_int32_round_to_zero( float128 );
    261 !!!int64 float128_to_int64( float128 );
    262 !!!int64 float128_to_int64_round_to_zero( float128 );
    263 float32 float128_to_float32( float128 );
    264 float64 float128_to_float64( float128 );
    265 #ifdef FLOATX80
    266 floatx80 float128_to_floatx80( float128 );
    267 #endif
    268 
    269 /*
    270 -------------------------------------------------------------------------------
    271 Software IEC/IEEE quadruple-precision operations.
    272 -------------------------------------------------------------------------------
    273 */
    274 float128 float128_round_to_int( float128 );
    275 float128 float128_add( float128, float128 );
    276 float128 float128_sub( float128, float128 );
    277 float128 float128_mul( float128, float128 );
    278 float128 float128_div( float128, float128 );
    279 float128 float128_rem( float128, float128 );
    280 float128 float128_sqrt( float128 );
    281 !!!flag float128_eq( float128, float128 );
    282 !!!flag float128_le( float128, float128 );
    283 !!!flag float128_lt( float128, float128 );
    284 !!!flag float128_eq_signaling( float128, float128 );
    285 !!!flag float128_le_quiet( float128, float128 );
    286 !!!flag float128_lt_quiet( float128, float128 );
    287 !!!flag float128_is_signaling_nan( float128 );
    288 
    289 #endif
    290 
    291