root / fpu / softfloat.c @ a63b5829
History | View | Annotate | Download (202 kB)
| 1 |
|
|---|---|
| 2 |
/*============================================================================
|
| 3 |
|
| 4 |
This C source file is part of the SoftFloat IEC/IEEE Floating-point Arithmetic
|
| 5 |
Package, Release 2b.
|
| 6 |
|
| 7 |
Written by John R. Hauser. This work was made possible in part by the
|
| 8 |
International Computer Science Institute, located at Suite 600, 1947 Center
|
| 9 |
Street, Berkeley, California 94704. Funding was partially provided by the
|
| 10 |
National Science Foundation under grant MIP-9311980. The original version
|
| 11 |
of this code was written as part of a project to build a fixed-point vector
|
| 12 |
processor in collaboration with the University of California at Berkeley,
|
| 13 |
overseen by Profs. Nelson Morgan and John Wawrzynek. More information
|
| 14 |
is available through the Web page `http://www.cs.berkeley.edu/~jhauser/
|
| 15 |
arithmetic/SoftFloat.html'.
|
| 16 |
|
| 17 |
THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE. Although reasonable effort has
|
| 18 |
been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT TIMES
|
| 19 |
RESULT IN INCORRECT BEHAVIOR. USE OF THIS SOFTWARE IS RESTRICTED TO PERSONS
|
| 20 |
AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ALL LOSSES,
|
| 21 |
COSTS, OR OTHER PROBLEMS THEY INCUR DUE TO THE SOFTWARE, AND WHO FURTHERMORE
|
| 22 |
EFFECTIVELY INDEMNIFY JOHN HAUSER AND THE INTERNATIONAL COMPUTER SCIENCE
|
| 23 |
INSTITUTE (possibly via similar legal warning) AGAINST ALL LOSSES, COSTS, OR
|
| 24 |
OTHER PROBLEMS INCURRED BY THEIR CUSTOMERS AND CLIENTS DUE TO THE SOFTWARE.
|
| 25 |
|
| 26 |
Derivative works are acceptable, even for commercial purposes, so long as
|
| 27 |
(1) the source code for the derivative work includes prominent notice that
|
| 28 |
the work is derivative, and (2) the source code includes prominent notice with
|
| 29 |
these four paragraphs for those parts of this code that are retained.
|
| 30 |
|
| 31 |
=============================================================================*/
|
| 32 |
|
| 33 |
/* FIXME: Flush-To-Zero only effects results. Denormal inputs should also
|
| 34 |
be flushed to zero. */
|
| 35 |
#include "softfloat.h" |
| 36 |
|
| 37 |
/*----------------------------------------------------------------------------
|
| 38 |
| Primitive arithmetic functions, including multi-word arithmetic, and
|
| 39 |
| division and square root approximations. (Can be specialized to target if
|
| 40 |
| desired.)
|
| 41 |
*----------------------------------------------------------------------------*/
|
| 42 |
#include "softfloat-macros.h" |
| 43 |
|
| 44 |
/*----------------------------------------------------------------------------
|
| 45 |
| Functions and definitions to determine: (1) whether tininess for underflow
|
| 46 |
| is detected before or after rounding by default, (2) what (if anything)
|
| 47 |
| happens when exceptions are raised, (3) how signaling NaNs are distinguished
|
| 48 |
| from quiet NaNs, (4) the default generated quiet NaNs, and (5) how NaNs
|
| 49 |
| are propagated from function inputs to output. These details are target-
|
| 50 |
| specific.
|
| 51 |
*----------------------------------------------------------------------------*/
|
| 52 |
#include "softfloat-specialize.h" |
| 53 |
|
| 54 |
void set_float_rounding_mode(int val STATUS_PARAM) |
| 55 |
{
|
| 56 |
STATUS(float_rounding_mode) = val; |
| 57 |
} |
| 58 |
|
| 59 |
void set_float_exception_flags(int val STATUS_PARAM) |
| 60 |
{
|
| 61 |
STATUS(float_exception_flags) = val; |
| 62 |
} |
| 63 |
|
| 64 |
#ifdef FLOATX80
|
| 65 |
void set_floatx80_rounding_precision(int val STATUS_PARAM) |
| 66 |
{
|
| 67 |
STATUS(floatx80_rounding_precision) = val; |
| 68 |
} |
| 69 |
#endif
|
| 70 |
|
| 71 |
/*----------------------------------------------------------------------------
|
| 72 |
| Takes a 64-bit fixed-point value `absZ' with binary point between bits 6
|
| 73 |
| and 7, and returns the properly rounded 32-bit integer corresponding to the
|
| 74 |
| input. If `zSign' is 1, the input is negated before being converted to an
|
| 75 |
| integer. Bit 63 of `absZ' must be zero. Ordinarily, the fixed-point input
|
| 76 |
| is simply rounded to an integer, with the inexact exception raised if the
|
| 77 |
| input cannot be represented exactly as an integer. However, if the fixed-
|
| 78 |
| point input is too large, the invalid exception is raised and the largest
|
| 79 |
| positive or negative integer is returned.
|
| 80 |
*----------------------------------------------------------------------------*/
|
| 81 |
|
| 82 |
static int32 roundAndPackInt32( flag zSign, bits64 absZ STATUS_PARAM)
|
| 83 |
{
|
| 84 |
int8 roundingMode; |
| 85 |
flag roundNearestEven; |
| 86 |
int8 roundIncrement, roundBits; |
| 87 |
int32 z; |
| 88 |
|
| 89 |
roundingMode = STATUS(float_rounding_mode); |
| 90 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 91 |
roundIncrement = 0x40;
|
| 92 |
if ( ! roundNearestEven ) {
|
| 93 |
if ( roundingMode == float_round_to_zero ) {
|
| 94 |
roundIncrement = 0;
|
| 95 |
} |
| 96 |
else {
|
| 97 |
roundIncrement = 0x7F;
|
| 98 |
if ( zSign ) {
|
| 99 |
if ( roundingMode == float_round_up ) roundIncrement = 0; |
| 100 |
} |
| 101 |
else {
|
| 102 |
if ( roundingMode == float_round_down ) roundIncrement = 0; |
| 103 |
} |
| 104 |
} |
| 105 |
} |
| 106 |
roundBits = absZ & 0x7F;
|
| 107 |
absZ = ( absZ + roundIncrement )>>7;
|
| 108 |
absZ &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); |
| 109 |
z = absZ; |
| 110 |
if ( zSign ) z = - z;
|
| 111 |
if ( ( absZ>>32 ) || ( z && ( ( z < 0 ) ^ zSign ) ) ) { |
| 112 |
float_raise( float_flag_invalid STATUS_VAR); |
| 113 |
return zSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; |
| 114 |
} |
| 115 |
if ( roundBits ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 116 |
return z;
|
| 117 |
|
| 118 |
} |
| 119 |
|
| 120 |
/*----------------------------------------------------------------------------
|
| 121 |
| Takes the 128-bit fixed-point value formed by concatenating `absZ0' and
|
| 122 |
| `absZ1', with binary point between bits 63 and 64 (between the input words),
|
| 123 |
| and returns the properly rounded 64-bit integer corresponding to the input.
|
| 124 |
| If `zSign' is 1, the input is negated before being converted to an integer.
|
| 125 |
| Ordinarily, the fixed-point input is simply rounded to an integer, with
|
| 126 |
| the inexact exception raised if the input cannot be represented exactly as
|
| 127 |
| an integer. However, if the fixed-point input is too large, the invalid
|
| 128 |
| exception is raised and the largest positive or negative integer is
|
| 129 |
| returned.
|
| 130 |
*----------------------------------------------------------------------------*/
|
| 131 |
|
| 132 |
static int64 roundAndPackInt64( flag zSign, bits64 absZ0, bits64 absZ1 STATUS_PARAM)
|
| 133 |
{
|
| 134 |
int8 roundingMode; |
| 135 |
flag roundNearestEven, increment; |
| 136 |
int64 z; |
| 137 |
|
| 138 |
roundingMode = STATUS(float_rounding_mode); |
| 139 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 140 |
increment = ( (sbits64) absZ1 < 0 );
|
| 141 |
if ( ! roundNearestEven ) {
|
| 142 |
if ( roundingMode == float_round_to_zero ) {
|
| 143 |
increment = 0;
|
| 144 |
} |
| 145 |
else {
|
| 146 |
if ( zSign ) {
|
| 147 |
increment = ( roundingMode == float_round_down ) && absZ1; |
| 148 |
} |
| 149 |
else {
|
| 150 |
increment = ( roundingMode == float_round_up ) && absZ1; |
| 151 |
} |
| 152 |
} |
| 153 |
} |
| 154 |
if ( increment ) {
|
| 155 |
++absZ0; |
| 156 |
if ( absZ0 == 0 ) goto overflow; |
| 157 |
absZ0 &= ~ ( ( (bits64) ( absZ1<<1 ) == 0 ) & roundNearestEven ); |
| 158 |
} |
| 159 |
z = absZ0; |
| 160 |
if ( zSign ) z = - z;
|
| 161 |
if ( z && ( ( z < 0 ) ^ zSign ) ) { |
| 162 |
overflow:
|
| 163 |
float_raise( float_flag_invalid STATUS_VAR); |
| 164 |
return
|
| 165 |
zSign ? (sbits64) LIT64( 0x8000000000000000 )
|
| 166 |
: LIT64( 0x7FFFFFFFFFFFFFFF );
|
| 167 |
} |
| 168 |
if ( absZ1 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 169 |
return z;
|
| 170 |
|
| 171 |
} |
| 172 |
|
| 173 |
/*----------------------------------------------------------------------------
|
| 174 |
| Returns the fraction bits of the single-precision floating-point value `a'.
|
| 175 |
*----------------------------------------------------------------------------*/
|
| 176 |
|
| 177 |
INLINE bits32 extractFloat32Frac( float32 a ) |
| 178 |
{
|
| 179 |
|
| 180 |
return float32_val(a) & 0x007FFFFF; |
| 181 |
|
| 182 |
} |
| 183 |
|
| 184 |
/*----------------------------------------------------------------------------
|
| 185 |
| Returns the exponent bits of the single-precision floating-point value `a'.
|
| 186 |
*----------------------------------------------------------------------------*/
|
| 187 |
|
| 188 |
INLINE int16 extractFloat32Exp( float32 a ) |
| 189 |
{
|
| 190 |
|
| 191 |
return ( float32_val(a)>>23 ) & 0xFF; |
| 192 |
|
| 193 |
} |
| 194 |
|
| 195 |
/*----------------------------------------------------------------------------
|
| 196 |
| Returns the sign bit of the single-precision floating-point value `a'.
|
| 197 |
*----------------------------------------------------------------------------*/
|
| 198 |
|
| 199 |
INLINE flag extractFloat32Sign( float32 a ) |
| 200 |
{
|
| 201 |
|
| 202 |
return float32_val(a)>>31; |
| 203 |
|
| 204 |
} |
| 205 |
|
| 206 |
/*----------------------------------------------------------------------------
|
| 207 |
| Normalizes the subnormal single-precision floating-point value represented
|
| 208 |
| by the denormalized significand `aSig'. The normalized exponent and
|
| 209 |
| significand are stored at the locations pointed to by `zExpPtr' and
|
| 210 |
| `zSigPtr', respectively.
|
| 211 |
*----------------------------------------------------------------------------*/
|
| 212 |
|
| 213 |
static void |
| 214 |
normalizeFloat32Subnormal( bits32 aSig, int16 *zExpPtr, bits32 *zSigPtr ) |
| 215 |
{
|
| 216 |
int8 shiftCount; |
| 217 |
|
| 218 |
shiftCount = countLeadingZeros32( aSig ) - 8;
|
| 219 |
*zSigPtr = aSig<<shiftCount; |
| 220 |
*zExpPtr = 1 - shiftCount;
|
| 221 |
|
| 222 |
} |
| 223 |
|
| 224 |
/*----------------------------------------------------------------------------
|
| 225 |
| Packs the sign `zSign', exponent `zExp', and significand `zSig' into a
|
| 226 |
| single-precision floating-point value, returning the result. After being
|
| 227 |
| shifted into the proper positions, the three fields are simply added
|
| 228 |
| together to form the result. This means that any integer portion of `zSig'
|
| 229 |
| will be added into the exponent. Since a properly normalized significand
|
| 230 |
| will have an integer portion equal to 1, the `zExp' input should be 1 less
|
| 231 |
| than the desired result exponent whenever `zSig' is a complete, normalized
|
| 232 |
| significand.
|
| 233 |
*----------------------------------------------------------------------------*/
|
| 234 |
|
| 235 |
INLINE float32 packFloat32( flag zSign, int16 zExp, bits32 zSig ) |
| 236 |
{
|
| 237 |
|
| 238 |
return make_float32(
|
| 239 |
( ( (bits32) zSign )<<31 ) + ( ( (bits32) zExp )<<23 ) + zSig); |
| 240 |
|
| 241 |
} |
| 242 |
|
| 243 |
/*----------------------------------------------------------------------------
|
| 244 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 245 |
| and significand `zSig', and returns the proper single-precision floating-
|
| 246 |
| point value corresponding to the abstract input. Ordinarily, the abstract
|
| 247 |
| value is simply rounded and packed into the single-precision format, with
|
| 248 |
| the inexact exception raised if the abstract input cannot be represented
|
| 249 |
| exactly. However, if the abstract value is too large, the overflow and
|
| 250 |
| inexact exceptions are raised and an infinity or maximal finite value is
|
| 251 |
| returned. If the abstract value is too small, the input value is rounded to
|
| 252 |
| a subnormal number, and the underflow and inexact exceptions are raised if
|
| 253 |
| the abstract input cannot be represented exactly as a subnormal single-
|
| 254 |
| precision floating-point number.
|
| 255 |
| The input significand `zSig' has its binary point between bits 30
|
| 256 |
| and 29, which is 7 bits to the left of the usual location. This shifted
|
| 257 |
| significand must be normalized or smaller. If `zSig' is not normalized,
|
| 258 |
| `zExp' must be 0; in that case, the result returned is a subnormal number,
|
| 259 |
| and it must not require rounding. In the usual case that `zSig' is
|
| 260 |
| normalized, `zExp' must be 1 less than the ``true'' floating-point exponent.
|
| 261 |
| The handling of underflow and overflow follows the IEC/IEEE Standard for
|
| 262 |
| Binary Floating-Point Arithmetic.
|
| 263 |
*----------------------------------------------------------------------------*/
|
| 264 |
|
| 265 |
static float32 roundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig STATUS_PARAM)
|
| 266 |
{
|
| 267 |
int8 roundingMode; |
| 268 |
flag roundNearestEven; |
| 269 |
int8 roundIncrement, roundBits; |
| 270 |
flag isTiny; |
| 271 |
|
| 272 |
roundingMode = STATUS(float_rounding_mode); |
| 273 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 274 |
roundIncrement = 0x40;
|
| 275 |
if ( ! roundNearestEven ) {
|
| 276 |
if ( roundingMode == float_round_to_zero ) {
|
| 277 |
roundIncrement = 0;
|
| 278 |
} |
| 279 |
else {
|
| 280 |
roundIncrement = 0x7F;
|
| 281 |
if ( zSign ) {
|
| 282 |
if ( roundingMode == float_round_up ) roundIncrement = 0; |
| 283 |
} |
| 284 |
else {
|
| 285 |
if ( roundingMode == float_round_down ) roundIncrement = 0; |
| 286 |
} |
| 287 |
} |
| 288 |
} |
| 289 |
roundBits = zSig & 0x7F;
|
| 290 |
if ( 0xFD <= (bits16) zExp ) { |
| 291 |
if ( ( 0xFD < zExp ) |
| 292 |
|| ( ( zExp == 0xFD )
|
| 293 |
&& ( (sbits32) ( zSig + roundIncrement ) < 0 ) )
|
| 294 |
) {
|
| 295 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 296 |
return packFloat32( zSign, 0xFF, - ( roundIncrement == 0 )); |
| 297 |
} |
| 298 |
if ( zExp < 0 ) { |
| 299 |
if ( STATUS(flush_to_zero) ) return packFloat32( zSign, 0, 0 ); |
| 300 |
isTiny = |
| 301 |
( STATUS(float_detect_tininess) == float_tininess_before_rounding ) |
| 302 |
|| ( zExp < -1 )
|
| 303 |
|| ( zSig + roundIncrement < 0x80000000 );
|
| 304 |
shift32RightJamming( zSig, - zExp, &zSig ); |
| 305 |
zExp = 0;
|
| 306 |
roundBits = zSig & 0x7F;
|
| 307 |
if ( isTiny && roundBits ) float_raise( float_flag_underflow STATUS_VAR);
|
| 308 |
} |
| 309 |
} |
| 310 |
if ( roundBits ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 311 |
zSig = ( zSig + roundIncrement )>>7;
|
| 312 |
zSig &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); |
| 313 |
if ( zSig == 0 ) zExp = 0; |
| 314 |
return packFloat32( zSign, zExp, zSig );
|
| 315 |
|
| 316 |
} |
| 317 |
|
| 318 |
/*----------------------------------------------------------------------------
|
| 319 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 320 |
| and significand `zSig', and returns the proper single-precision floating-
|
| 321 |
| point value corresponding to the abstract input. This routine is just like
|
| 322 |
| `roundAndPackFloat32' except that `zSig' does not have to be normalized.
|
| 323 |
| Bit 31 of `zSig' must be zero, and `zExp' must be 1 less than the ``true''
|
| 324 |
| floating-point exponent.
|
| 325 |
*----------------------------------------------------------------------------*/
|
| 326 |
|
| 327 |
static float32
|
| 328 |
normalizeRoundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig STATUS_PARAM) |
| 329 |
{
|
| 330 |
int8 shiftCount; |
| 331 |
|
| 332 |
shiftCount = countLeadingZeros32( zSig ) - 1;
|
| 333 |
return roundAndPackFloat32( zSign, zExp - shiftCount, zSig<<shiftCount STATUS_VAR);
|
| 334 |
|
| 335 |
} |
| 336 |
|
| 337 |
/*----------------------------------------------------------------------------
|
| 338 |
| Returns the fraction bits of the double-precision floating-point value `a'.
|
| 339 |
*----------------------------------------------------------------------------*/
|
| 340 |
|
| 341 |
INLINE bits64 extractFloat64Frac( float64 a ) |
| 342 |
{
|
| 343 |
|
| 344 |
return float64_val(a) & LIT64( 0x000FFFFFFFFFFFFF ); |
| 345 |
|
| 346 |
} |
| 347 |
|
| 348 |
/*----------------------------------------------------------------------------
|
| 349 |
| Returns the exponent bits of the double-precision floating-point value `a'.
|
| 350 |
*----------------------------------------------------------------------------*/
|
| 351 |
|
| 352 |
INLINE int16 extractFloat64Exp( float64 a ) |
| 353 |
{
|
| 354 |
|
| 355 |
return ( float64_val(a)>>52 ) & 0x7FF; |
| 356 |
|
| 357 |
} |
| 358 |
|
| 359 |
/*----------------------------------------------------------------------------
|
| 360 |
| Returns the sign bit of the double-precision floating-point value `a'.
|
| 361 |
*----------------------------------------------------------------------------*/
|
| 362 |
|
| 363 |
INLINE flag extractFloat64Sign( float64 a ) |
| 364 |
{
|
| 365 |
|
| 366 |
return float64_val(a)>>63; |
| 367 |
|
| 368 |
} |
| 369 |
|
| 370 |
/*----------------------------------------------------------------------------
|
| 371 |
| Normalizes the subnormal double-precision floating-point value represented
|
| 372 |
| by the denormalized significand `aSig'. The normalized exponent and
|
| 373 |
| significand are stored at the locations pointed to by `zExpPtr' and
|
| 374 |
| `zSigPtr', respectively.
|
| 375 |
*----------------------------------------------------------------------------*/
|
| 376 |
|
| 377 |
static void |
| 378 |
normalizeFloat64Subnormal( bits64 aSig, int16 *zExpPtr, bits64 *zSigPtr ) |
| 379 |
{
|
| 380 |
int8 shiftCount; |
| 381 |
|
| 382 |
shiftCount = countLeadingZeros64( aSig ) - 11;
|
| 383 |
*zSigPtr = aSig<<shiftCount; |
| 384 |
*zExpPtr = 1 - shiftCount;
|
| 385 |
|
| 386 |
} |
| 387 |
|
| 388 |
/*----------------------------------------------------------------------------
|
| 389 |
| Packs the sign `zSign', exponent `zExp', and significand `zSig' into a
|
| 390 |
| double-precision floating-point value, returning the result. After being
|
| 391 |
| shifted into the proper positions, the three fields are simply added
|
| 392 |
| together to form the result. This means that any integer portion of `zSig'
|
| 393 |
| will be added into the exponent. Since a properly normalized significand
|
| 394 |
| will have an integer portion equal to 1, the `zExp' input should be 1 less
|
| 395 |
| than the desired result exponent whenever `zSig' is a complete, normalized
|
| 396 |
| significand.
|
| 397 |
*----------------------------------------------------------------------------*/
|
| 398 |
|
| 399 |
INLINE float64 packFloat64( flag zSign, int16 zExp, bits64 zSig ) |
| 400 |
{
|
| 401 |
|
| 402 |
return make_float64(
|
| 403 |
( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<52 ) + zSig); |
| 404 |
|
| 405 |
} |
| 406 |
|
| 407 |
/*----------------------------------------------------------------------------
|
| 408 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 409 |
| and significand `zSig', and returns the proper double-precision floating-
|
| 410 |
| point value corresponding to the abstract input. Ordinarily, the abstract
|
| 411 |
| value is simply rounded and packed into the double-precision format, with
|
| 412 |
| the inexact exception raised if the abstract input cannot be represented
|
| 413 |
| exactly. However, if the abstract value is too large, the overflow and
|
| 414 |
| inexact exceptions are raised and an infinity or maximal finite value is
|
| 415 |
| returned. If the abstract value is too small, the input value is rounded
|
| 416 |
| to a subnormal number, and the underflow and inexact exceptions are raised
|
| 417 |
| if the abstract input cannot be represented exactly as a subnormal double-
|
| 418 |
| precision floating-point number.
|
| 419 |
| The input significand `zSig' has its binary point between bits 62
|
| 420 |
| and 61, which is 10 bits to the left of the usual location. This shifted
|
| 421 |
| significand must be normalized or smaller. If `zSig' is not normalized,
|
| 422 |
| `zExp' must be 0; in that case, the result returned is a subnormal number,
|
| 423 |
| and it must not require rounding. In the usual case that `zSig' is
|
| 424 |
| normalized, `zExp' must be 1 less than the ``true'' floating-point exponent.
|
| 425 |
| The handling of underflow and overflow follows the IEC/IEEE Standard for
|
| 426 |
| Binary Floating-Point Arithmetic.
|
| 427 |
*----------------------------------------------------------------------------*/
|
| 428 |
|
| 429 |
static float64 roundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig STATUS_PARAM)
|
| 430 |
{
|
| 431 |
int8 roundingMode; |
| 432 |
flag roundNearestEven; |
| 433 |
int16 roundIncrement, roundBits; |
| 434 |
flag isTiny; |
| 435 |
|
| 436 |
roundingMode = STATUS(float_rounding_mode); |
| 437 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 438 |
roundIncrement = 0x200;
|
| 439 |
if ( ! roundNearestEven ) {
|
| 440 |
if ( roundingMode == float_round_to_zero ) {
|
| 441 |
roundIncrement = 0;
|
| 442 |
} |
| 443 |
else {
|
| 444 |
roundIncrement = 0x3FF;
|
| 445 |
if ( zSign ) {
|
| 446 |
if ( roundingMode == float_round_up ) roundIncrement = 0; |
| 447 |
} |
| 448 |
else {
|
| 449 |
if ( roundingMode == float_round_down ) roundIncrement = 0; |
| 450 |
} |
| 451 |
} |
| 452 |
} |
| 453 |
roundBits = zSig & 0x3FF;
|
| 454 |
if ( 0x7FD <= (bits16) zExp ) { |
| 455 |
if ( ( 0x7FD < zExp ) |
| 456 |
|| ( ( zExp == 0x7FD )
|
| 457 |
&& ( (sbits64) ( zSig + roundIncrement ) < 0 ) )
|
| 458 |
) {
|
| 459 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 460 |
return packFloat64( zSign, 0x7FF, - ( roundIncrement == 0 )); |
| 461 |
} |
| 462 |
if ( zExp < 0 ) { |
| 463 |
if ( STATUS(flush_to_zero) ) return packFloat64( zSign, 0, 0 ); |
| 464 |
isTiny = |
| 465 |
( STATUS(float_detect_tininess) == float_tininess_before_rounding ) |
| 466 |
|| ( zExp < -1 )
|
| 467 |
|| ( zSig + roundIncrement < LIT64( 0x8000000000000000 ) );
|
| 468 |
shift64RightJamming( zSig, - zExp, &zSig ); |
| 469 |
zExp = 0;
|
| 470 |
roundBits = zSig & 0x3FF;
|
| 471 |
if ( isTiny && roundBits ) float_raise( float_flag_underflow STATUS_VAR);
|
| 472 |
} |
| 473 |
} |
| 474 |
if ( roundBits ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 475 |
zSig = ( zSig + roundIncrement )>>10;
|
| 476 |
zSig &= ~ ( ( ( roundBits ^ 0x200 ) == 0 ) & roundNearestEven ); |
| 477 |
if ( zSig == 0 ) zExp = 0; |
| 478 |
return packFloat64( zSign, zExp, zSig );
|
| 479 |
|
| 480 |
} |
| 481 |
|
| 482 |
/*----------------------------------------------------------------------------
|
| 483 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 484 |
| and significand `zSig', and returns the proper double-precision floating-
|
| 485 |
| point value corresponding to the abstract input. This routine is just like
|
| 486 |
| `roundAndPackFloat64' except that `zSig' does not have to be normalized.
|
| 487 |
| Bit 63 of `zSig' must be zero, and `zExp' must be 1 less than the ``true''
|
| 488 |
| floating-point exponent.
|
| 489 |
*----------------------------------------------------------------------------*/
|
| 490 |
|
| 491 |
static float64
|
| 492 |
normalizeRoundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig STATUS_PARAM) |
| 493 |
{
|
| 494 |
int8 shiftCount; |
| 495 |
|
| 496 |
shiftCount = countLeadingZeros64( zSig ) - 1;
|
| 497 |
return roundAndPackFloat64( zSign, zExp - shiftCount, zSig<<shiftCount STATUS_VAR);
|
| 498 |
|
| 499 |
} |
| 500 |
|
| 501 |
#ifdef FLOATX80
|
| 502 |
|
| 503 |
/*----------------------------------------------------------------------------
|
| 504 |
| Returns the fraction bits of the extended double-precision floating-point
|
| 505 |
| value `a'.
|
| 506 |
*----------------------------------------------------------------------------*/
|
| 507 |
|
| 508 |
INLINE bits64 extractFloatx80Frac( floatx80 a ) |
| 509 |
{
|
| 510 |
|
| 511 |
return a.low;
|
| 512 |
|
| 513 |
} |
| 514 |
|
| 515 |
/*----------------------------------------------------------------------------
|
| 516 |
| Returns the exponent bits of the extended double-precision floating-point
|
| 517 |
| value `a'.
|
| 518 |
*----------------------------------------------------------------------------*/
|
| 519 |
|
| 520 |
INLINE int32 extractFloatx80Exp( floatx80 a ) |
| 521 |
{
|
| 522 |
|
| 523 |
return a.high & 0x7FFF; |
| 524 |
|
| 525 |
} |
| 526 |
|
| 527 |
/*----------------------------------------------------------------------------
|
| 528 |
| Returns the sign bit of the extended double-precision floating-point value
|
| 529 |
| `a'.
|
| 530 |
*----------------------------------------------------------------------------*/
|
| 531 |
|
| 532 |
INLINE flag extractFloatx80Sign( floatx80 a ) |
| 533 |
{
|
| 534 |
|
| 535 |
return a.high>>15; |
| 536 |
|
| 537 |
} |
| 538 |
|
| 539 |
/*----------------------------------------------------------------------------
|
| 540 |
| Normalizes the subnormal extended double-precision floating-point value
|
| 541 |
| represented by the denormalized significand `aSig'. The normalized exponent
|
| 542 |
| and significand are stored at the locations pointed to by `zExpPtr' and
|
| 543 |
| `zSigPtr', respectively.
|
| 544 |
*----------------------------------------------------------------------------*/
|
| 545 |
|
| 546 |
static void |
| 547 |
normalizeFloatx80Subnormal( bits64 aSig, int32 *zExpPtr, bits64 *zSigPtr ) |
| 548 |
{
|
| 549 |
int8 shiftCount; |
| 550 |
|
| 551 |
shiftCount = countLeadingZeros64( aSig ); |
| 552 |
*zSigPtr = aSig<<shiftCount; |
| 553 |
*zExpPtr = 1 - shiftCount;
|
| 554 |
|
| 555 |
} |
| 556 |
|
| 557 |
/*----------------------------------------------------------------------------
|
| 558 |
| Packs the sign `zSign', exponent `zExp', and significand `zSig' into an
|
| 559 |
| extended double-precision floating-point value, returning the result.
|
| 560 |
*----------------------------------------------------------------------------*/
|
| 561 |
|
| 562 |
INLINE floatx80 packFloatx80( flag zSign, int32 zExp, bits64 zSig ) |
| 563 |
{
|
| 564 |
floatx80 z; |
| 565 |
|
| 566 |
z.low = zSig; |
| 567 |
z.high = ( ( (bits16) zSign )<<15 ) + zExp;
|
| 568 |
return z;
|
| 569 |
|
| 570 |
} |
| 571 |
|
| 572 |
/*----------------------------------------------------------------------------
|
| 573 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 574 |
| and extended significand formed by the concatenation of `zSig0' and `zSig1',
|
| 575 |
| and returns the proper extended double-precision floating-point value
|
| 576 |
| corresponding to the abstract input. Ordinarily, the abstract value is
|
| 577 |
| rounded and packed into the extended double-precision format, with the
|
| 578 |
| inexact exception raised if the abstract input cannot be represented
|
| 579 |
| exactly. However, if the abstract value is too large, the overflow and
|
| 580 |
| inexact exceptions are raised and an infinity or maximal finite value is
|
| 581 |
| returned. If the abstract value is too small, the input value is rounded to
|
| 582 |
| a subnormal number, and the underflow and inexact exceptions are raised if
|
| 583 |
| the abstract input cannot be represented exactly as a subnormal extended
|
| 584 |
| double-precision floating-point number.
|
| 585 |
| If `roundingPrecision' is 32 or 64, the result is rounded to the same
|
| 586 |
| number of bits as single or double precision, respectively. Otherwise, the
|
| 587 |
| result is rounded to the full precision of the extended double-precision
|
| 588 |
| format.
|
| 589 |
| The input significand must be normalized or smaller. If the input
|
| 590 |
| significand is not normalized, `zExp' must be 0; in that case, the result
|
| 591 |
| returned is a subnormal number, and it must not require rounding. The
|
| 592 |
| handling of underflow and overflow follows the IEC/IEEE Standard for Binary
|
| 593 |
| Floating-Point Arithmetic.
|
| 594 |
*----------------------------------------------------------------------------*/
|
| 595 |
|
| 596 |
static floatx80
|
| 597 |
roundAndPackFloatx80( |
| 598 |
int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 |
| 599 |
STATUS_PARAM) |
| 600 |
{
|
| 601 |
int8 roundingMode; |
| 602 |
flag roundNearestEven, increment, isTiny; |
| 603 |
int64 roundIncrement, roundMask, roundBits; |
| 604 |
|
| 605 |
roundingMode = STATUS(float_rounding_mode); |
| 606 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 607 |
if ( roundingPrecision == 80 ) goto precision80; |
| 608 |
if ( roundingPrecision == 64 ) { |
| 609 |
roundIncrement = LIT64( 0x0000000000000400 );
|
| 610 |
roundMask = LIT64( 0x00000000000007FF );
|
| 611 |
} |
| 612 |
else if ( roundingPrecision == 32 ) { |
| 613 |
roundIncrement = LIT64( 0x0000008000000000 );
|
| 614 |
roundMask = LIT64( 0x000000FFFFFFFFFF );
|
| 615 |
} |
| 616 |
else {
|
| 617 |
goto precision80;
|
| 618 |
} |
| 619 |
zSig0 |= ( zSig1 != 0 );
|
| 620 |
if ( ! roundNearestEven ) {
|
| 621 |
if ( roundingMode == float_round_to_zero ) {
|
| 622 |
roundIncrement = 0;
|
| 623 |
} |
| 624 |
else {
|
| 625 |
roundIncrement = roundMask; |
| 626 |
if ( zSign ) {
|
| 627 |
if ( roundingMode == float_round_up ) roundIncrement = 0; |
| 628 |
} |
| 629 |
else {
|
| 630 |
if ( roundingMode == float_round_down ) roundIncrement = 0; |
| 631 |
} |
| 632 |
} |
| 633 |
} |
| 634 |
roundBits = zSig0 & roundMask; |
| 635 |
if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { |
| 636 |
if ( ( 0x7FFE < zExp ) |
| 637 |
|| ( ( zExp == 0x7FFE ) && ( zSig0 + roundIncrement < zSig0 ) )
|
| 638 |
) {
|
| 639 |
goto overflow;
|
| 640 |
} |
| 641 |
if ( zExp <= 0 ) { |
| 642 |
if ( STATUS(flush_to_zero) ) return packFloatx80( zSign, 0, 0 ); |
| 643 |
isTiny = |
| 644 |
( STATUS(float_detect_tininess) == float_tininess_before_rounding ) |
| 645 |
|| ( zExp < 0 )
|
| 646 |
|| ( zSig0 <= zSig0 + roundIncrement ); |
| 647 |
shift64RightJamming( zSig0, 1 - zExp, &zSig0 );
|
| 648 |
zExp = 0;
|
| 649 |
roundBits = zSig0 & roundMask; |
| 650 |
if ( isTiny && roundBits ) float_raise( float_flag_underflow STATUS_VAR);
|
| 651 |
if ( roundBits ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 652 |
zSig0 += roundIncrement; |
| 653 |
if ( (sbits64) zSig0 < 0 ) zExp = 1; |
| 654 |
roundIncrement = roundMask + 1;
|
| 655 |
if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { |
| 656 |
roundMask |= roundIncrement; |
| 657 |
} |
| 658 |
zSig0 &= ~ roundMask; |
| 659 |
return packFloatx80( zSign, zExp, zSig0 );
|
| 660 |
} |
| 661 |
} |
| 662 |
if ( roundBits ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 663 |
zSig0 += roundIncrement; |
| 664 |
if ( zSig0 < roundIncrement ) {
|
| 665 |
++zExp; |
| 666 |
zSig0 = LIT64( 0x8000000000000000 );
|
| 667 |
} |
| 668 |
roundIncrement = roundMask + 1;
|
| 669 |
if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { |
| 670 |
roundMask |= roundIncrement; |
| 671 |
} |
| 672 |
zSig0 &= ~ roundMask; |
| 673 |
if ( zSig0 == 0 ) zExp = 0; |
| 674 |
return packFloatx80( zSign, zExp, zSig0 );
|
| 675 |
precision80:
|
| 676 |
increment = ( (sbits64) zSig1 < 0 );
|
| 677 |
if ( ! roundNearestEven ) {
|
| 678 |
if ( roundingMode == float_round_to_zero ) {
|
| 679 |
increment = 0;
|
| 680 |
} |
| 681 |
else {
|
| 682 |
if ( zSign ) {
|
| 683 |
increment = ( roundingMode == float_round_down ) && zSig1; |
| 684 |
} |
| 685 |
else {
|
| 686 |
increment = ( roundingMode == float_round_up ) && zSig1; |
| 687 |
} |
| 688 |
} |
| 689 |
} |
| 690 |
if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { |
| 691 |
if ( ( 0x7FFE < zExp ) |
| 692 |
|| ( ( zExp == 0x7FFE )
|
| 693 |
&& ( zSig0 == LIT64( 0xFFFFFFFFFFFFFFFF ) )
|
| 694 |
&& increment |
| 695 |
) |
| 696 |
) {
|
| 697 |
roundMask = 0;
|
| 698 |
overflow:
|
| 699 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 700 |
if ( ( roundingMode == float_round_to_zero )
|
| 701 |
|| ( zSign && ( roundingMode == float_round_up ) ) |
| 702 |
|| ( ! zSign && ( roundingMode == float_round_down ) ) |
| 703 |
) {
|
| 704 |
return packFloatx80( zSign, 0x7FFE, ~ roundMask ); |
| 705 |
} |
| 706 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 707 |
} |
| 708 |
if ( zExp <= 0 ) { |
| 709 |
isTiny = |
| 710 |
( STATUS(float_detect_tininess) == float_tininess_before_rounding ) |
| 711 |
|| ( zExp < 0 )
|
| 712 |
|| ! increment |
| 713 |
|| ( zSig0 < LIT64( 0xFFFFFFFFFFFFFFFF ) );
|
| 714 |
shift64ExtraRightJamming( zSig0, zSig1, 1 - zExp, &zSig0, &zSig1 );
|
| 715 |
zExp = 0;
|
| 716 |
if ( isTiny && zSig1 ) float_raise( float_flag_underflow STATUS_VAR);
|
| 717 |
if ( zSig1 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 718 |
if ( roundNearestEven ) {
|
| 719 |
increment = ( (sbits64) zSig1 < 0 );
|
| 720 |
} |
| 721 |
else {
|
| 722 |
if ( zSign ) {
|
| 723 |
increment = ( roundingMode == float_round_down ) && zSig1; |
| 724 |
} |
| 725 |
else {
|
| 726 |
increment = ( roundingMode == float_round_up ) && zSig1; |
| 727 |
} |
| 728 |
} |
| 729 |
if ( increment ) {
|
| 730 |
++zSig0; |
| 731 |
zSig0 &= |
| 732 |
~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); |
| 733 |
if ( (sbits64) zSig0 < 0 ) zExp = 1; |
| 734 |
} |
| 735 |
return packFloatx80( zSign, zExp, zSig0 );
|
| 736 |
} |
| 737 |
} |
| 738 |
if ( zSig1 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 739 |
if ( increment ) {
|
| 740 |
++zSig0; |
| 741 |
if ( zSig0 == 0 ) { |
| 742 |
++zExp; |
| 743 |
zSig0 = LIT64( 0x8000000000000000 );
|
| 744 |
} |
| 745 |
else {
|
| 746 |
zSig0 &= ~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); |
| 747 |
} |
| 748 |
} |
| 749 |
else {
|
| 750 |
if ( zSig0 == 0 ) zExp = 0; |
| 751 |
} |
| 752 |
return packFloatx80( zSign, zExp, zSig0 );
|
| 753 |
|
| 754 |
} |
| 755 |
|
| 756 |
/*----------------------------------------------------------------------------
|
| 757 |
| Takes an abstract floating-point value having sign `zSign', exponent
|
| 758 |
| `zExp', and significand formed by the concatenation of `zSig0' and `zSig1',
|
| 759 |
| and returns the proper extended double-precision floating-point value
|
| 760 |
| corresponding to the abstract input. This routine is just like
|
| 761 |
| `roundAndPackFloatx80' except that the input significand does not have to be
|
| 762 |
| normalized.
|
| 763 |
*----------------------------------------------------------------------------*/
|
| 764 |
|
| 765 |
static floatx80
|
| 766 |
normalizeRoundAndPackFloatx80( |
| 767 |
int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 |
| 768 |
STATUS_PARAM) |
| 769 |
{
|
| 770 |
int8 shiftCount; |
| 771 |
|
| 772 |
if ( zSig0 == 0 ) { |
| 773 |
zSig0 = zSig1; |
| 774 |
zSig1 = 0;
|
| 775 |
zExp -= 64;
|
| 776 |
} |
| 777 |
shiftCount = countLeadingZeros64( zSig0 ); |
| 778 |
shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); |
| 779 |
zExp -= shiftCount; |
| 780 |
return
|
| 781 |
roundAndPackFloatx80( roundingPrecision, zSign, zExp, zSig0, zSig1 STATUS_VAR); |
| 782 |
|
| 783 |
} |
| 784 |
|
| 785 |
#endif
|
| 786 |
|
| 787 |
#ifdef FLOAT128
|
| 788 |
|
| 789 |
/*----------------------------------------------------------------------------
|
| 790 |
| Returns the least-significant 64 fraction bits of the quadruple-precision
|
| 791 |
| floating-point value `a'.
|
| 792 |
*----------------------------------------------------------------------------*/
|
| 793 |
|
| 794 |
INLINE bits64 extractFloat128Frac1( float128 a ) |
| 795 |
{
|
| 796 |
|
| 797 |
return a.low;
|
| 798 |
|
| 799 |
} |
| 800 |
|
| 801 |
/*----------------------------------------------------------------------------
|
| 802 |
| Returns the most-significant 48 fraction bits of the quadruple-precision
|
| 803 |
| floating-point value `a'.
|
| 804 |
*----------------------------------------------------------------------------*/
|
| 805 |
|
| 806 |
INLINE bits64 extractFloat128Frac0( float128 a ) |
| 807 |
{
|
| 808 |
|
| 809 |
return a.high & LIT64( 0x0000FFFFFFFFFFFF ); |
| 810 |
|
| 811 |
} |
| 812 |
|
| 813 |
/*----------------------------------------------------------------------------
|
| 814 |
| Returns the exponent bits of the quadruple-precision floating-point value
|
| 815 |
| `a'.
|
| 816 |
*----------------------------------------------------------------------------*/
|
| 817 |
|
| 818 |
INLINE int32 extractFloat128Exp( float128 a ) |
| 819 |
{
|
| 820 |
|
| 821 |
return ( a.high>>48 ) & 0x7FFF; |
| 822 |
|
| 823 |
} |
| 824 |
|
| 825 |
/*----------------------------------------------------------------------------
|
| 826 |
| Returns the sign bit of the quadruple-precision floating-point value `a'.
|
| 827 |
*----------------------------------------------------------------------------*/
|
| 828 |
|
| 829 |
INLINE flag extractFloat128Sign( float128 a ) |
| 830 |
{
|
| 831 |
|
| 832 |
return a.high>>63; |
| 833 |
|
| 834 |
} |
| 835 |
|
| 836 |
/*----------------------------------------------------------------------------
|
| 837 |
| Normalizes the subnormal quadruple-precision floating-point value
|
| 838 |
| represented by the denormalized significand formed by the concatenation of
|
| 839 |
| `aSig0' and `aSig1'. The normalized exponent is stored at the location
|
| 840 |
| pointed to by `zExpPtr'. The most significant 49 bits of the normalized
|
| 841 |
| significand are stored at the location pointed to by `zSig0Ptr', and the
|
| 842 |
| least significant 64 bits of the normalized significand are stored at the
|
| 843 |
| location pointed to by `zSig1Ptr'.
|
| 844 |
*----------------------------------------------------------------------------*/
|
| 845 |
|
| 846 |
static void |
| 847 |
normalizeFloat128Subnormal( |
| 848 |
bits64 aSig0, |
| 849 |
bits64 aSig1, |
| 850 |
int32 *zExpPtr, |
| 851 |
bits64 *zSig0Ptr, |
| 852 |
bits64 *zSig1Ptr |
| 853 |
) |
| 854 |
{
|
| 855 |
int8 shiftCount; |
| 856 |
|
| 857 |
if ( aSig0 == 0 ) { |
| 858 |
shiftCount = countLeadingZeros64( aSig1 ) - 15;
|
| 859 |
if ( shiftCount < 0 ) { |
| 860 |
*zSig0Ptr = aSig1>>( - shiftCount ); |
| 861 |
*zSig1Ptr = aSig1<<( shiftCount & 63 );
|
| 862 |
} |
| 863 |
else {
|
| 864 |
*zSig0Ptr = aSig1<<shiftCount; |
| 865 |
*zSig1Ptr = 0;
|
| 866 |
} |
| 867 |
*zExpPtr = - shiftCount - 63;
|
| 868 |
} |
| 869 |
else {
|
| 870 |
shiftCount = countLeadingZeros64( aSig0 ) - 15;
|
| 871 |
shortShift128Left( aSig0, aSig1, shiftCount, zSig0Ptr, zSig1Ptr ); |
| 872 |
*zExpPtr = 1 - shiftCount;
|
| 873 |
} |
| 874 |
|
| 875 |
} |
| 876 |
|
| 877 |
/*----------------------------------------------------------------------------
|
| 878 |
| Packs the sign `zSign', the exponent `zExp', and the significand formed
|
| 879 |
| by the concatenation of `zSig0' and `zSig1' into a quadruple-precision
|
| 880 |
| floating-point value, returning the result. After being shifted into the
|
| 881 |
| proper positions, the three fields `zSign', `zExp', and `zSig0' are simply
|
| 882 |
| added together to form the most significant 32 bits of the result. This
|
| 883 |
| means that any integer portion of `zSig0' will be added into the exponent.
|
| 884 |
| Since a properly normalized significand will have an integer portion equal
|
| 885 |
| to 1, the `zExp' input should be 1 less than the desired result exponent
|
| 886 |
| whenever `zSig0' and `zSig1' concatenated form a complete, normalized
|
| 887 |
| significand.
|
| 888 |
*----------------------------------------------------------------------------*/
|
| 889 |
|
| 890 |
INLINE float128 |
| 891 |
packFloat128( flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 ) |
| 892 |
{
|
| 893 |
float128 z; |
| 894 |
|
| 895 |
z.low = zSig1; |
| 896 |
z.high = ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<48 ) + zSig0; |
| 897 |
return z;
|
| 898 |
|
| 899 |
} |
| 900 |
|
| 901 |
/*----------------------------------------------------------------------------
|
| 902 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 903 |
| and extended significand formed by the concatenation of `zSig0', `zSig1',
|
| 904 |
| and `zSig2', and returns the proper quadruple-precision floating-point value
|
| 905 |
| corresponding to the abstract input. Ordinarily, the abstract value is
|
| 906 |
| simply rounded and packed into the quadruple-precision format, with the
|
| 907 |
| inexact exception raised if the abstract input cannot be represented
|
| 908 |
| exactly. However, if the abstract value is too large, the overflow and
|
| 909 |
| inexact exceptions are raised and an infinity or maximal finite value is
|
| 910 |
| returned. If the abstract value is too small, the input value is rounded to
|
| 911 |
| a subnormal number, and the underflow and inexact exceptions are raised if
|
| 912 |
| the abstract input cannot be represented exactly as a subnormal quadruple-
|
| 913 |
| precision floating-point number.
|
| 914 |
| The input significand must be normalized or smaller. If the input
|
| 915 |
| significand is not normalized, `zExp' must be 0; in that case, the result
|
| 916 |
| returned is a subnormal number, and it must not require rounding. In the
|
| 917 |
| usual case that the input significand is normalized, `zExp' must be 1 less
|
| 918 |
| than the ``true'' floating-point exponent. The handling of underflow and
|
| 919 |
| overflow follows the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 920 |
*----------------------------------------------------------------------------*/
|
| 921 |
|
| 922 |
static float128
|
| 923 |
roundAndPackFloat128( |
| 924 |
flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1, bits64 zSig2 STATUS_PARAM) |
| 925 |
{
|
| 926 |
int8 roundingMode; |
| 927 |
flag roundNearestEven, increment, isTiny; |
| 928 |
|
| 929 |
roundingMode = STATUS(float_rounding_mode); |
| 930 |
roundNearestEven = ( roundingMode == float_round_nearest_even ); |
| 931 |
increment = ( (sbits64) zSig2 < 0 );
|
| 932 |
if ( ! roundNearestEven ) {
|
| 933 |
if ( roundingMode == float_round_to_zero ) {
|
| 934 |
increment = 0;
|
| 935 |
} |
| 936 |
else {
|
| 937 |
if ( zSign ) {
|
| 938 |
increment = ( roundingMode == float_round_down ) && zSig2; |
| 939 |
} |
| 940 |
else {
|
| 941 |
increment = ( roundingMode == float_round_up ) && zSig2; |
| 942 |
} |
| 943 |
} |
| 944 |
} |
| 945 |
if ( 0x7FFD <= (bits32) zExp ) { |
| 946 |
if ( ( 0x7FFD < zExp ) |
| 947 |
|| ( ( zExp == 0x7FFD )
|
| 948 |
&& eq128( |
| 949 |
LIT64( 0x0001FFFFFFFFFFFF ),
|
| 950 |
LIT64( 0xFFFFFFFFFFFFFFFF ),
|
| 951 |
zSig0, |
| 952 |
zSig1 |
| 953 |
) |
| 954 |
&& increment |
| 955 |
) |
| 956 |
) {
|
| 957 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 958 |
if ( ( roundingMode == float_round_to_zero )
|
| 959 |
|| ( zSign && ( roundingMode == float_round_up ) ) |
| 960 |
|| ( ! zSign && ( roundingMode == float_round_down ) ) |
| 961 |
) {
|
| 962 |
return
|
| 963 |
packFloat128( |
| 964 |
zSign, |
| 965 |
0x7FFE,
|
| 966 |
LIT64( 0x0000FFFFFFFFFFFF ),
|
| 967 |
LIT64( 0xFFFFFFFFFFFFFFFF )
|
| 968 |
); |
| 969 |
} |
| 970 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 971 |
} |
| 972 |
if ( zExp < 0 ) { |
| 973 |
if ( STATUS(flush_to_zero) ) return packFloat128( zSign, 0, 0, 0 ); |
| 974 |
isTiny = |
| 975 |
( STATUS(float_detect_tininess) == float_tininess_before_rounding ) |
| 976 |
|| ( zExp < -1 )
|
| 977 |
|| ! increment |
| 978 |
|| lt128( |
| 979 |
zSig0, |
| 980 |
zSig1, |
| 981 |
LIT64( 0x0001FFFFFFFFFFFF ),
|
| 982 |
LIT64( 0xFFFFFFFFFFFFFFFF )
|
| 983 |
); |
| 984 |
shift128ExtraRightJamming( |
| 985 |
zSig0, zSig1, zSig2, - zExp, &zSig0, &zSig1, &zSig2 ); |
| 986 |
zExp = 0;
|
| 987 |
if ( isTiny && zSig2 ) float_raise( float_flag_underflow STATUS_VAR);
|
| 988 |
if ( roundNearestEven ) {
|
| 989 |
increment = ( (sbits64) zSig2 < 0 );
|
| 990 |
} |
| 991 |
else {
|
| 992 |
if ( zSign ) {
|
| 993 |
increment = ( roundingMode == float_round_down ) && zSig2; |
| 994 |
} |
| 995 |
else {
|
| 996 |
increment = ( roundingMode == float_round_up ) && zSig2; |
| 997 |
} |
| 998 |
} |
| 999 |
} |
| 1000 |
} |
| 1001 |
if ( zSig2 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 1002 |
if ( increment ) {
|
| 1003 |
add128( zSig0, zSig1, 0, 1, &zSig0, &zSig1 ); |
| 1004 |
zSig1 &= ~ ( ( zSig2 + zSig2 == 0 ) & roundNearestEven );
|
| 1005 |
} |
| 1006 |
else {
|
| 1007 |
if ( ( zSig0 | zSig1 ) == 0 ) zExp = 0; |
| 1008 |
} |
| 1009 |
return packFloat128( zSign, zExp, zSig0, zSig1 );
|
| 1010 |
|
| 1011 |
} |
| 1012 |
|
| 1013 |
/*----------------------------------------------------------------------------
|
| 1014 |
| Takes an abstract floating-point value having sign `zSign', exponent `zExp',
|
| 1015 |
| and significand formed by the concatenation of `zSig0' and `zSig1', and
|
| 1016 |
| returns the proper quadruple-precision floating-point value corresponding
|
| 1017 |
| to the abstract input. This routine is just like `roundAndPackFloat128'
|
| 1018 |
| except that the input significand has fewer bits and does not have to be
|
| 1019 |
| normalized. In all cases, `zExp' must be 1 less than the ``true'' floating-
|
| 1020 |
| point exponent.
|
| 1021 |
*----------------------------------------------------------------------------*/
|
| 1022 |
|
| 1023 |
static float128
|
| 1024 |
normalizeRoundAndPackFloat128( |
| 1025 |
flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 STATUS_PARAM) |
| 1026 |
{
|
| 1027 |
int8 shiftCount; |
| 1028 |
bits64 zSig2; |
| 1029 |
|
| 1030 |
if ( zSig0 == 0 ) { |
| 1031 |
zSig0 = zSig1; |
| 1032 |
zSig1 = 0;
|
| 1033 |
zExp -= 64;
|
| 1034 |
} |
| 1035 |
shiftCount = countLeadingZeros64( zSig0 ) - 15;
|
| 1036 |
if ( 0 <= shiftCount ) { |
| 1037 |
zSig2 = 0;
|
| 1038 |
shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); |
| 1039 |
} |
| 1040 |
else {
|
| 1041 |
shift128ExtraRightJamming( |
| 1042 |
zSig0, zSig1, 0, - shiftCount, &zSig0, &zSig1, &zSig2 );
|
| 1043 |
} |
| 1044 |
zExp -= shiftCount; |
| 1045 |
return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 STATUS_VAR);
|
| 1046 |
|
| 1047 |
} |
| 1048 |
|
| 1049 |
#endif
|
| 1050 |
|
| 1051 |
/*----------------------------------------------------------------------------
|
| 1052 |
| Returns the result of converting the 32-bit two's complement integer `a'
|
| 1053 |
| to the single-precision floating-point format. The conversion is performed
|
| 1054 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1055 |
*----------------------------------------------------------------------------*/
|
| 1056 |
|
| 1057 |
float32 int32_to_float32( int32 a STATUS_PARAM ) |
| 1058 |
{
|
| 1059 |
flag zSign; |
| 1060 |
|
| 1061 |
if ( a == 0 ) return float32_zero; |
| 1062 |
if ( a == (sbits32) 0x80000000 ) return packFloat32( 1, 0x9E, 0 ); |
| 1063 |
zSign = ( a < 0 );
|
| 1064 |
return normalizeRoundAndPackFloat32( zSign, 0x9C, zSign ? - a : a STATUS_VAR ); |
| 1065 |
|
| 1066 |
} |
| 1067 |
|
| 1068 |
/*----------------------------------------------------------------------------
|
| 1069 |
| Returns the result of converting the 32-bit two's complement integer `a'
|
| 1070 |
| to the double-precision floating-point format. The conversion is performed
|
| 1071 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1072 |
*----------------------------------------------------------------------------*/
|
| 1073 |
|
| 1074 |
float64 int32_to_float64( int32 a STATUS_PARAM ) |
| 1075 |
{
|
| 1076 |
flag zSign; |
| 1077 |
uint32 absA; |
| 1078 |
int8 shiftCount; |
| 1079 |
bits64 zSig; |
| 1080 |
|
| 1081 |
if ( a == 0 ) return float64_zero; |
| 1082 |
zSign = ( a < 0 );
|
| 1083 |
absA = zSign ? - a : a; |
| 1084 |
shiftCount = countLeadingZeros32( absA ) + 21;
|
| 1085 |
zSig = absA; |
| 1086 |
return packFloat64( zSign, 0x432 - shiftCount, zSig<<shiftCount ); |
| 1087 |
|
| 1088 |
} |
| 1089 |
|
| 1090 |
#ifdef FLOATX80
|
| 1091 |
|
| 1092 |
/*----------------------------------------------------------------------------
|
| 1093 |
| Returns the result of converting the 32-bit two's complement integer `a'
|
| 1094 |
| to the extended double-precision floating-point format. The conversion
|
| 1095 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1096 |
| Arithmetic.
|
| 1097 |
*----------------------------------------------------------------------------*/
|
| 1098 |
|
| 1099 |
floatx80 int32_to_floatx80( int32 a STATUS_PARAM ) |
| 1100 |
{
|
| 1101 |
flag zSign; |
| 1102 |
uint32 absA; |
| 1103 |
int8 shiftCount; |
| 1104 |
bits64 zSig; |
| 1105 |
|
| 1106 |
if ( a == 0 ) return packFloatx80( 0, 0, 0 ); |
| 1107 |
zSign = ( a < 0 );
|
| 1108 |
absA = zSign ? - a : a; |
| 1109 |
shiftCount = countLeadingZeros32( absA ) + 32;
|
| 1110 |
zSig = absA; |
| 1111 |
return packFloatx80( zSign, 0x403E - shiftCount, zSig<<shiftCount ); |
| 1112 |
|
| 1113 |
} |
| 1114 |
|
| 1115 |
#endif
|
| 1116 |
|
| 1117 |
#ifdef FLOAT128
|
| 1118 |
|
| 1119 |
/*----------------------------------------------------------------------------
|
| 1120 |
| Returns the result of converting the 32-bit two's complement integer `a' to
|
| 1121 |
| the quadruple-precision floating-point format. The conversion is performed
|
| 1122 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1123 |
*----------------------------------------------------------------------------*/
|
| 1124 |
|
| 1125 |
float128 int32_to_float128( int32 a STATUS_PARAM ) |
| 1126 |
{
|
| 1127 |
flag zSign; |
| 1128 |
uint32 absA; |
| 1129 |
int8 shiftCount; |
| 1130 |
bits64 zSig0; |
| 1131 |
|
| 1132 |
if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); |
| 1133 |
zSign = ( a < 0 );
|
| 1134 |
absA = zSign ? - a : a; |
| 1135 |
shiftCount = countLeadingZeros32( absA ) + 17;
|
| 1136 |
zSig0 = absA; |
| 1137 |
return packFloat128( zSign, 0x402E - shiftCount, zSig0<<shiftCount, 0 ); |
| 1138 |
|
| 1139 |
} |
| 1140 |
|
| 1141 |
#endif
|
| 1142 |
|
| 1143 |
/*----------------------------------------------------------------------------
|
| 1144 |
| Returns the result of converting the 64-bit two's complement integer `a'
|
| 1145 |
| to the single-precision floating-point format. The conversion is performed
|
| 1146 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1147 |
*----------------------------------------------------------------------------*/
|
| 1148 |
|
| 1149 |
float32 int64_to_float32( int64 a STATUS_PARAM ) |
| 1150 |
{
|
| 1151 |
flag zSign; |
| 1152 |
uint64 absA; |
| 1153 |
int8 shiftCount; |
| 1154 |
|
| 1155 |
if ( a == 0 ) return float32_zero; |
| 1156 |
zSign = ( a < 0 );
|
| 1157 |
absA = zSign ? - a : a; |
| 1158 |
shiftCount = countLeadingZeros64( absA ) - 40;
|
| 1159 |
if ( 0 <= shiftCount ) { |
| 1160 |
return packFloat32( zSign, 0x95 - shiftCount, absA<<shiftCount ); |
| 1161 |
} |
| 1162 |
else {
|
| 1163 |
shiftCount += 7;
|
| 1164 |
if ( shiftCount < 0 ) { |
| 1165 |
shift64RightJamming( absA, - shiftCount, &absA ); |
| 1166 |
} |
| 1167 |
else {
|
| 1168 |
absA <<= shiftCount; |
| 1169 |
} |
| 1170 |
return roundAndPackFloat32( zSign, 0x9C - shiftCount, absA STATUS_VAR ); |
| 1171 |
} |
| 1172 |
|
| 1173 |
} |
| 1174 |
|
| 1175 |
float32 uint64_to_float32( uint64 a STATUS_PARAM ) |
| 1176 |
{
|
| 1177 |
int8 shiftCount; |
| 1178 |
|
| 1179 |
if ( a == 0 ) return float32_zero; |
| 1180 |
shiftCount = countLeadingZeros64( a ) - 40;
|
| 1181 |
if ( 0 <= shiftCount ) { |
| 1182 |
return packFloat32( 1 > 0, 0x95 - shiftCount, a<<shiftCount ); |
| 1183 |
} |
| 1184 |
else {
|
| 1185 |
shiftCount += 7;
|
| 1186 |
if ( shiftCount < 0 ) { |
| 1187 |
shift64RightJamming( a, - shiftCount, &a ); |
| 1188 |
} |
| 1189 |
else {
|
| 1190 |
a <<= shiftCount; |
| 1191 |
} |
| 1192 |
return roundAndPackFloat32( 1 > 0, 0x9C - shiftCount, a STATUS_VAR ); |
| 1193 |
} |
| 1194 |
} |
| 1195 |
|
| 1196 |
/*----------------------------------------------------------------------------
|
| 1197 |
| Returns the result of converting the 64-bit two's complement integer `a'
|
| 1198 |
| to the double-precision floating-point format. The conversion is performed
|
| 1199 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1200 |
*----------------------------------------------------------------------------*/
|
| 1201 |
|
| 1202 |
float64 int64_to_float64( int64 a STATUS_PARAM ) |
| 1203 |
{
|
| 1204 |
flag zSign; |
| 1205 |
|
| 1206 |
if ( a == 0 ) return float64_zero; |
| 1207 |
if ( a == (sbits64) LIT64( 0x8000000000000000 ) ) { |
| 1208 |
return packFloat64( 1, 0x43E, 0 ); |
| 1209 |
} |
| 1210 |
zSign = ( a < 0 );
|
| 1211 |
return normalizeRoundAndPackFloat64( zSign, 0x43C, zSign ? - a : a STATUS_VAR ); |
| 1212 |
|
| 1213 |
} |
| 1214 |
|
| 1215 |
float64 uint64_to_float64( uint64 a STATUS_PARAM ) |
| 1216 |
{
|
| 1217 |
if ( a == 0 ) return float64_zero; |
| 1218 |
return normalizeRoundAndPackFloat64( 0, 0x43C, a STATUS_VAR ); |
| 1219 |
|
| 1220 |
} |
| 1221 |
|
| 1222 |
#ifdef FLOATX80
|
| 1223 |
|
| 1224 |
/*----------------------------------------------------------------------------
|
| 1225 |
| Returns the result of converting the 64-bit two's complement integer `a'
|
| 1226 |
| to the extended double-precision floating-point format. The conversion
|
| 1227 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1228 |
| Arithmetic.
|
| 1229 |
*----------------------------------------------------------------------------*/
|
| 1230 |
|
| 1231 |
floatx80 int64_to_floatx80( int64 a STATUS_PARAM ) |
| 1232 |
{
|
| 1233 |
flag zSign; |
| 1234 |
uint64 absA; |
| 1235 |
int8 shiftCount; |
| 1236 |
|
| 1237 |
if ( a == 0 ) return packFloatx80( 0, 0, 0 ); |
| 1238 |
zSign = ( a < 0 );
|
| 1239 |
absA = zSign ? - a : a; |
| 1240 |
shiftCount = countLeadingZeros64( absA ); |
| 1241 |
return packFloatx80( zSign, 0x403E - shiftCount, absA<<shiftCount ); |
| 1242 |
|
| 1243 |
} |
| 1244 |
|
| 1245 |
#endif
|
| 1246 |
|
| 1247 |
#ifdef FLOAT128
|
| 1248 |
|
| 1249 |
/*----------------------------------------------------------------------------
|
| 1250 |
| Returns the result of converting the 64-bit two's complement integer `a' to
|
| 1251 |
| the quadruple-precision floating-point format. The conversion is performed
|
| 1252 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1253 |
*----------------------------------------------------------------------------*/
|
| 1254 |
|
| 1255 |
float128 int64_to_float128( int64 a STATUS_PARAM ) |
| 1256 |
{
|
| 1257 |
flag zSign; |
| 1258 |
uint64 absA; |
| 1259 |
int8 shiftCount; |
| 1260 |
int32 zExp; |
| 1261 |
bits64 zSig0, zSig1; |
| 1262 |
|
| 1263 |
if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); |
| 1264 |
zSign = ( a < 0 );
|
| 1265 |
absA = zSign ? - a : a; |
| 1266 |
shiftCount = countLeadingZeros64( absA ) + 49;
|
| 1267 |
zExp = 0x406E - shiftCount;
|
| 1268 |
if ( 64 <= shiftCount ) { |
| 1269 |
zSig1 = 0;
|
| 1270 |
zSig0 = absA; |
| 1271 |
shiftCount -= 64;
|
| 1272 |
} |
| 1273 |
else {
|
| 1274 |
zSig1 = absA; |
| 1275 |
zSig0 = 0;
|
| 1276 |
} |
| 1277 |
shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); |
| 1278 |
return packFloat128( zSign, zExp, zSig0, zSig1 );
|
| 1279 |
|
| 1280 |
} |
| 1281 |
|
| 1282 |
#endif
|
| 1283 |
|
| 1284 |
/*----------------------------------------------------------------------------
|
| 1285 |
| Returns the result of converting the single-precision floating-point value
|
| 1286 |
| `a' to the 32-bit two's complement integer format. The conversion is
|
| 1287 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1288 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 1289 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 1290 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 1291 |
| largest integer with the same sign as `a' is returned.
|
| 1292 |
*----------------------------------------------------------------------------*/
|
| 1293 |
|
| 1294 |
int32 float32_to_int32( float32 a STATUS_PARAM ) |
| 1295 |
{
|
| 1296 |
flag aSign; |
| 1297 |
int16 aExp, shiftCount; |
| 1298 |
bits32 aSig; |
| 1299 |
bits64 aSig64; |
| 1300 |
|
| 1301 |
aSig = extractFloat32Frac( a ); |
| 1302 |
aExp = extractFloat32Exp( a ); |
| 1303 |
aSign = extractFloat32Sign( a ); |
| 1304 |
if ( ( aExp == 0xFF ) && aSig ) aSign = 0; |
| 1305 |
if ( aExp ) aSig |= 0x00800000; |
| 1306 |
shiftCount = 0xAF - aExp;
|
| 1307 |
aSig64 = aSig; |
| 1308 |
aSig64 <<= 32;
|
| 1309 |
if ( 0 < shiftCount ) shift64RightJamming( aSig64, shiftCount, &aSig64 ); |
| 1310 |
return roundAndPackInt32( aSign, aSig64 STATUS_VAR );
|
| 1311 |
|
| 1312 |
} |
| 1313 |
|
| 1314 |
/*----------------------------------------------------------------------------
|
| 1315 |
| Returns the result of converting the single-precision floating-point value
|
| 1316 |
| `a' to the 32-bit two's complement integer format. The conversion is
|
| 1317 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1318 |
| Arithmetic, except that the conversion is always rounded toward zero.
|
| 1319 |
| If `a' is a NaN, the largest positive integer is returned. Otherwise, if
|
| 1320 |
| the conversion overflows, the largest integer with the same sign as `a' is
|
| 1321 |
| returned.
|
| 1322 |
*----------------------------------------------------------------------------*/
|
| 1323 |
|
| 1324 |
int32 float32_to_int32_round_to_zero( float32 a STATUS_PARAM ) |
| 1325 |
{
|
| 1326 |
flag aSign; |
| 1327 |
int16 aExp, shiftCount; |
| 1328 |
bits32 aSig; |
| 1329 |
int32 z; |
| 1330 |
|
| 1331 |
aSig = extractFloat32Frac( a ); |
| 1332 |
aExp = extractFloat32Exp( a ); |
| 1333 |
aSign = extractFloat32Sign( a ); |
| 1334 |
shiftCount = aExp - 0x9E;
|
| 1335 |
if ( 0 <= shiftCount ) { |
| 1336 |
if ( float32_val(a) != 0xCF000000 ) { |
| 1337 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1338 |
if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) return 0x7FFFFFFF; |
| 1339 |
} |
| 1340 |
return (sbits32) 0x80000000; |
| 1341 |
} |
| 1342 |
else if ( aExp <= 0x7E ) { |
| 1343 |
if ( aExp | aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 1344 |
return 0; |
| 1345 |
} |
| 1346 |
aSig = ( aSig | 0x00800000 )<<8; |
| 1347 |
z = aSig>>( - shiftCount ); |
| 1348 |
if ( (bits32) ( aSig<<( shiftCount & 31 ) ) ) { |
| 1349 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 1350 |
} |
| 1351 |
if ( aSign ) z = - z;
|
| 1352 |
return z;
|
| 1353 |
|
| 1354 |
} |
| 1355 |
|
| 1356 |
/*----------------------------------------------------------------------------
|
| 1357 |
| Returns the result of converting the single-precision floating-point value
|
| 1358 |
| `a' to the 64-bit two's complement integer format. The conversion is
|
| 1359 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1360 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 1361 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 1362 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 1363 |
| largest integer with the same sign as `a' is returned.
|
| 1364 |
*----------------------------------------------------------------------------*/
|
| 1365 |
|
| 1366 |
int64 float32_to_int64( float32 a STATUS_PARAM ) |
| 1367 |
{
|
| 1368 |
flag aSign; |
| 1369 |
int16 aExp, shiftCount; |
| 1370 |
bits32 aSig; |
| 1371 |
bits64 aSig64, aSigExtra; |
| 1372 |
|
| 1373 |
aSig = extractFloat32Frac( a ); |
| 1374 |
aExp = extractFloat32Exp( a ); |
| 1375 |
aSign = extractFloat32Sign( a ); |
| 1376 |
shiftCount = 0xBE - aExp;
|
| 1377 |
if ( shiftCount < 0 ) { |
| 1378 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1379 |
if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { |
| 1380 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 1381 |
} |
| 1382 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 1383 |
} |
| 1384 |
if ( aExp ) aSig |= 0x00800000; |
| 1385 |
aSig64 = aSig; |
| 1386 |
aSig64 <<= 40;
|
| 1387 |
shift64ExtraRightJamming( aSig64, 0, shiftCount, &aSig64, &aSigExtra );
|
| 1388 |
return roundAndPackInt64( aSign, aSig64, aSigExtra STATUS_VAR );
|
| 1389 |
|
| 1390 |
} |
| 1391 |
|
| 1392 |
/*----------------------------------------------------------------------------
|
| 1393 |
| Returns the result of converting the single-precision floating-point value
|
| 1394 |
| `a' to the 64-bit two's complement integer format. The conversion is
|
| 1395 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1396 |
| Arithmetic, except that the conversion is always rounded toward zero. If
|
| 1397 |
| `a' is a NaN, the largest positive integer is returned. Otherwise, if the
|
| 1398 |
| conversion overflows, the largest integer with the same sign as `a' is
|
| 1399 |
| returned.
|
| 1400 |
*----------------------------------------------------------------------------*/
|
| 1401 |
|
| 1402 |
int64 float32_to_int64_round_to_zero( float32 a STATUS_PARAM ) |
| 1403 |
{
|
| 1404 |
flag aSign; |
| 1405 |
int16 aExp, shiftCount; |
| 1406 |
bits32 aSig; |
| 1407 |
bits64 aSig64; |
| 1408 |
int64 z; |
| 1409 |
|
| 1410 |
aSig = extractFloat32Frac( a ); |
| 1411 |
aExp = extractFloat32Exp( a ); |
| 1412 |
aSign = extractFloat32Sign( a ); |
| 1413 |
shiftCount = aExp - 0xBE;
|
| 1414 |
if ( 0 <= shiftCount ) { |
| 1415 |
if ( float32_val(a) != 0xDF000000 ) { |
| 1416 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1417 |
if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { |
| 1418 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 1419 |
} |
| 1420 |
} |
| 1421 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 1422 |
} |
| 1423 |
else if ( aExp <= 0x7E ) { |
| 1424 |
if ( aExp | aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 1425 |
return 0; |
| 1426 |
} |
| 1427 |
aSig64 = aSig | 0x00800000;
|
| 1428 |
aSig64 <<= 40;
|
| 1429 |
z = aSig64>>( - shiftCount ); |
| 1430 |
if ( (bits64) ( aSig64<<( shiftCount & 63 ) ) ) { |
| 1431 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 1432 |
} |
| 1433 |
if ( aSign ) z = - z;
|
| 1434 |
return z;
|
| 1435 |
|
| 1436 |
} |
| 1437 |
|
| 1438 |
/*----------------------------------------------------------------------------
|
| 1439 |
| Returns the result of converting the single-precision floating-point value
|
| 1440 |
| `a' to the double-precision floating-point format. The conversion is
|
| 1441 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1442 |
| Arithmetic.
|
| 1443 |
*----------------------------------------------------------------------------*/
|
| 1444 |
|
| 1445 |
float64 float32_to_float64( float32 a STATUS_PARAM ) |
| 1446 |
{
|
| 1447 |
flag aSign; |
| 1448 |
int16 aExp; |
| 1449 |
bits32 aSig; |
| 1450 |
|
| 1451 |
aSig = extractFloat32Frac( a ); |
| 1452 |
aExp = extractFloat32Exp( a ); |
| 1453 |
aSign = extractFloat32Sign( a ); |
| 1454 |
if ( aExp == 0xFF ) { |
| 1455 |
if ( aSig ) return commonNaNToFloat64( float32ToCommonNaN( a STATUS_VAR )); |
| 1456 |
return packFloat64( aSign, 0x7FF, 0 ); |
| 1457 |
} |
| 1458 |
if ( aExp == 0 ) { |
| 1459 |
if ( aSig == 0 ) return packFloat64( aSign, 0, 0 ); |
| 1460 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1461 |
--aExp; |
| 1462 |
} |
| 1463 |
return packFloat64( aSign, aExp + 0x380, ( (bits64) aSig )<<29 ); |
| 1464 |
|
| 1465 |
} |
| 1466 |
|
| 1467 |
#ifdef FLOATX80
|
| 1468 |
|
| 1469 |
/*----------------------------------------------------------------------------
|
| 1470 |
| Returns the result of converting the single-precision floating-point value
|
| 1471 |
| `a' to the extended double-precision floating-point format. The conversion
|
| 1472 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1473 |
| Arithmetic.
|
| 1474 |
*----------------------------------------------------------------------------*/
|
| 1475 |
|
| 1476 |
floatx80 float32_to_floatx80( float32 a STATUS_PARAM ) |
| 1477 |
{
|
| 1478 |
flag aSign; |
| 1479 |
int16 aExp; |
| 1480 |
bits32 aSig; |
| 1481 |
|
| 1482 |
aSig = extractFloat32Frac( a ); |
| 1483 |
aExp = extractFloat32Exp( a ); |
| 1484 |
aSign = extractFloat32Sign( a ); |
| 1485 |
if ( aExp == 0xFF ) { |
| 1486 |
if ( aSig ) return commonNaNToFloatx80( float32ToCommonNaN( a STATUS_VAR ) ); |
| 1487 |
return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 1488 |
} |
| 1489 |
if ( aExp == 0 ) { |
| 1490 |
if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); |
| 1491 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1492 |
} |
| 1493 |
aSig |= 0x00800000;
|
| 1494 |
return packFloatx80( aSign, aExp + 0x3F80, ( (bits64) aSig )<<40 ); |
| 1495 |
|
| 1496 |
} |
| 1497 |
|
| 1498 |
#endif
|
| 1499 |
|
| 1500 |
#ifdef FLOAT128
|
| 1501 |
|
| 1502 |
/*----------------------------------------------------------------------------
|
| 1503 |
| Returns the result of converting the single-precision floating-point value
|
| 1504 |
| `a' to the double-precision floating-point format. The conversion is
|
| 1505 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 1506 |
| Arithmetic.
|
| 1507 |
*----------------------------------------------------------------------------*/
|
| 1508 |
|
| 1509 |
float128 float32_to_float128( float32 a STATUS_PARAM ) |
| 1510 |
{
|
| 1511 |
flag aSign; |
| 1512 |
int16 aExp; |
| 1513 |
bits32 aSig; |
| 1514 |
|
| 1515 |
aSig = extractFloat32Frac( a ); |
| 1516 |
aExp = extractFloat32Exp( a ); |
| 1517 |
aSign = extractFloat32Sign( a ); |
| 1518 |
if ( aExp == 0xFF ) { |
| 1519 |
if ( aSig ) return commonNaNToFloat128( float32ToCommonNaN( a STATUS_VAR ) ); |
| 1520 |
return packFloat128( aSign, 0x7FFF, 0, 0 ); |
| 1521 |
} |
| 1522 |
if ( aExp == 0 ) { |
| 1523 |
if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); |
| 1524 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1525 |
--aExp; |
| 1526 |
} |
| 1527 |
return packFloat128( aSign, aExp + 0x3F80, ( (bits64) aSig )<<25, 0 ); |
| 1528 |
|
| 1529 |
} |
| 1530 |
|
| 1531 |
#endif
|
| 1532 |
|
| 1533 |
/*----------------------------------------------------------------------------
|
| 1534 |
| Rounds the single-precision floating-point value `a' to an integer, and
|
| 1535 |
| returns the result as a single-precision floating-point value. The
|
| 1536 |
| operation is performed according to the IEC/IEEE Standard for Binary
|
| 1537 |
| Floating-Point Arithmetic.
|
| 1538 |
*----------------------------------------------------------------------------*/
|
| 1539 |
|
| 1540 |
float32 float32_round_to_int( float32 a STATUS_PARAM) |
| 1541 |
{
|
| 1542 |
flag aSign; |
| 1543 |
int16 aExp; |
| 1544 |
bits32 lastBitMask, roundBitsMask; |
| 1545 |
int8 roundingMode; |
| 1546 |
bits32 z; |
| 1547 |
|
| 1548 |
aExp = extractFloat32Exp( a ); |
| 1549 |
if ( 0x96 <= aExp ) { |
| 1550 |
if ( ( aExp == 0xFF ) && extractFloat32Frac( a ) ) { |
| 1551 |
return propagateFloat32NaN( a, a STATUS_VAR );
|
| 1552 |
} |
| 1553 |
return a;
|
| 1554 |
} |
| 1555 |
if ( aExp <= 0x7E ) { |
| 1556 |
if ( (bits32) ( float32_val(a)<<1 ) == 0 ) return a; |
| 1557 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 1558 |
aSign = extractFloat32Sign( a ); |
| 1559 |
switch ( STATUS(float_rounding_mode) ) {
|
| 1560 |
case float_round_nearest_even:
|
| 1561 |
if ( ( aExp == 0x7E ) && extractFloat32Frac( a ) ) { |
| 1562 |
return packFloat32( aSign, 0x7F, 0 ); |
| 1563 |
} |
| 1564 |
break;
|
| 1565 |
case float_round_down:
|
| 1566 |
return make_float32(aSign ? 0xBF800000 : 0); |
| 1567 |
case float_round_up:
|
| 1568 |
return make_float32(aSign ? 0x80000000 : 0x3F800000); |
| 1569 |
} |
| 1570 |
return packFloat32( aSign, 0, 0 ); |
| 1571 |
} |
| 1572 |
lastBitMask = 1;
|
| 1573 |
lastBitMask <<= 0x96 - aExp;
|
| 1574 |
roundBitsMask = lastBitMask - 1;
|
| 1575 |
z = float32_val(a); |
| 1576 |
roundingMode = STATUS(float_rounding_mode); |
| 1577 |
if ( roundingMode == float_round_nearest_even ) {
|
| 1578 |
z += lastBitMask>>1;
|
| 1579 |
if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; |
| 1580 |
} |
| 1581 |
else if ( roundingMode != float_round_to_zero ) { |
| 1582 |
if ( extractFloat32Sign( make_float32(z) ) ^ ( roundingMode == float_round_up ) ) {
|
| 1583 |
z += roundBitsMask; |
| 1584 |
} |
| 1585 |
} |
| 1586 |
z &= ~ roundBitsMask; |
| 1587 |
if ( z != float32_val(a) ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 1588 |
return make_float32(z);
|
| 1589 |
|
| 1590 |
} |
| 1591 |
|
| 1592 |
/*----------------------------------------------------------------------------
|
| 1593 |
| Returns the result of adding the absolute values of the single-precision
|
| 1594 |
| floating-point values `a' and `b'. If `zSign' is 1, the sum is negated
|
| 1595 |
| before being returned. `zSign' is ignored if the result is a NaN.
|
| 1596 |
| The addition is performed according to the IEC/IEEE Standard for Binary
|
| 1597 |
| Floating-Point Arithmetic.
|
| 1598 |
*----------------------------------------------------------------------------*/
|
| 1599 |
|
| 1600 |
static float32 addFloat32Sigs( float32 a, float32 b, flag zSign STATUS_PARAM)
|
| 1601 |
{
|
| 1602 |
int16 aExp, bExp, zExp; |
| 1603 |
bits32 aSig, bSig, zSig; |
| 1604 |
int16 expDiff; |
| 1605 |
|
| 1606 |
aSig = extractFloat32Frac( a ); |
| 1607 |
aExp = extractFloat32Exp( a ); |
| 1608 |
bSig = extractFloat32Frac( b ); |
| 1609 |
bExp = extractFloat32Exp( b ); |
| 1610 |
expDiff = aExp - bExp; |
| 1611 |
aSig <<= 6;
|
| 1612 |
bSig <<= 6;
|
| 1613 |
if ( 0 < expDiff ) { |
| 1614 |
if ( aExp == 0xFF ) { |
| 1615 |
if ( aSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1616 |
return a;
|
| 1617 |
} |
| 1618 |
if ( bExp == 0 ) { |
| 1619 |
--expDiff; |
| 1620 |
} |
| 1621 |
else {
|
| 1622 |
bSig |= 0x20000000;
|
| 1623 |
} |
| 1624 |
shift32RightJamming( bSig, expDiff, &bSig ); |
| 1625 |
zExp = aExp; |
| 1626 |
} |
| 1627 |
else if ( expDiff < 0 ) { |
| 1628 |
if ( bExp == 0xFF ) { |
| 1629 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1630 |
return packFloat32( zSign, 0xFF, 0 ); |
| 1631 |
} |
| 1632 |
if ( aExp == 0 ) { |
| 1633 |
++expDiff; |
| 1634 |
} |
| 1635 |
else {
|
| 1636 |
aSig |= 0x20000000;
|
| 1637 |
} |
| 1638 |
shift32RightJamming( aSig, - expDiff, &aSig ); |
| 1639 |
zExp = bExp; |
| 1640 |
} |
| 1641 |
else {
|
| 1642 |
if ( aExp == 0xFF ) { |
| 1643 |
if ( aSig | bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1644 |
return a;
|
| 1645 |
} |
| 1646 |
if ( aExp == 0 ) { |
| 1647 |
if ( STATUS(flush_to_zero) ) return packFloat32( zSign, 0, 0 ); |
| 1648 |
return packFloat32( zSign, 0, ( aSig + bSig )>>6 ); |
| 1649 |
} |
| 1650 |
zSig = 0x40000000 + aSig + bSig;
|
| 1651 |
zExp = aExp; |
| 1652 |
goto roundAndPack;
|
| 1653 |
} |
| 1654 |
aSig |= 0x20000000;
|
| 1655 |
zSig = ( aSig + bSig )<<1;
|
| 1656 |
--zExp; |
| 1657 |
if ( (sbits32) zSig < 0 ) { |
| 1658 |
zSig = aSig + bSig; |
| 1659 |
++zExp; |
| 1660 |
} |
| 1661 |
roundAndPack:
|
| 1662 |
return roundAndPackFloat32( zSign, zExp, zSig STATUS_VAR );
|
| 1663 |
|
| 1664 |
} |
| 1665 |
|
| 1666 |
/*----------------------------------------------------------------------------
|
| 1667 |
| Returns the result of subtracting the absolute values of the single-
|
| 1668 |
| precision floating-point values `a' and `b'. If `zSign' is 1, the
|
| 1669 |
| difference is negated before being returned. `zSign' is ignored if the
|
| 1670 |
| result is a NaN. The subtraction is performed according to the IEC/IEEE
|
| 1671 |
| Standard for Binary Floating-Point Arithmetic.
|
| 1672 |
*----------------------------------------------------------------------------*/
|
| 1673 |
|
| 1674 |
static float32 subFloat32Sigs( float32 a, float32 b, flag zSign STATUS_PARAM)
|
| 1675 |
{
|
| 1676 |
int16 aExp, bExp, zExp; |
| 1677 |
bits32 aSig, bSig, zSig; |
| 1678 |
int16 expDiff; |
| 1679 |
|
| 1680 |
aSig = extractFloat32Frac( a ); |
| 1681 |
aExp = extractFloat32Exp( a ); |
| 1682 |
bSig = extractFloat32Frac( b ); |
| 1683 |
bExp = extractFloat32Exp( b ); |
| 1684 |
expDiff = aExp - bExp; |
| 1685 |
aSig <<= 7;
|
| 1686 |
bSig <<= 7;
|
| 1687 |
if ( 0 < expDiff ) goto aExpBigger; |
| 1688 |
if ( expDiff < 0 ) goto bExpBigger; |
| 1689 |
if ( aExp == 0xFF ) { |
| 1690 |
if ( aSig | bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1691 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1692 |
return float32_default_nan;
|
| 1693 |
} |
| 1694 |
if ( aExp == 0 ) { |
| 1695 |
aExp = 1;
|
| 1696 |
bExp = 1;
|
| 1697 |
} |
| 1698 |
if ( bSig < aSig ) goto aBigger; |
| 1699 |
if ( aSig < bSig ) goto bBigger; |
| 1700 |
return packFloat32( STATUS(float_rounding_mode) == float_round_down, 0, 0 ); |
| 1701 |
bExpBigger:
|
| 1702 |
if ( bExp == 0xFF ) { |
| 1703 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1704 |
return packFloat32( zSign ^ 1, 0xFF, 0 ); |
| 1705 |
} |
| 1706 |
if ( aExp == 0 ) { |
| 1707 |
++expDiff; |
| 1708 |
} |
| 1709 |
else {
|
| 1710 |
aSig |= 0x40000000;
|
| 1711 |
} |
| 1712 |
shift32RightJamming( aSig, - expDiff, &aSig ); |
| 1713 |
bSig |= 0x40000000;
|
| 1714 |
bBigger:
|
| 1715 |
zSig = bSig - aSig; |
| 1716 |
zExp = bExp; |
| 1717 |
zSign ^= 1;
|
| 1718 |
goto normalizeRoundAndPack;
|
| 1719 |
aExpBigger:
|
| 1720 |
if ( aExp == 0xFF ) { |
| 1721 |
if ( aSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1722 |
return a;
|
| 1723 |
} |
| 1724 |
if ( bExp == 0 ) { |
| 1725 |
--expDiff; |
| 1726 |
} |
| 1727 |
else {
|
| 1728 |
bSig |= 0x40000000;
|
| 1729 |
} |
| 1730 |
shift32RightJamming( bSig, expDiff, &bSig ); |
| 1731 |
aSig |= 0x40000000;
|
| 1732 |
aBigger:
|
| 1733 |
zSig = aSig - bSig; |
| 1734 |
zExp = aExp; |
| 1735 |
normalizeRoundAndPack:
|
| 1736 |
--zExp; |
| 1737 |
return normalizeRoundAndPackFloat32( zSign, zExp, zSig STATUS_VAR );
|
| 1738 |
|
| 1739 |
} |
| 1740 |
|
| 1741 |
/*----------------------------------------------------------------------------
|
| 1742 |
| Returns the result of adding the single-precision floating-point values `a'
|
| 1743 |
| and `b'. The operation is performed according to the IEC/IEEE Standard for
|
| 1744 |
| Binary Floating-Point Arithmetic.
|
| 1745 |
*----------------------------------------------------------------------------*/
|
| 1746 |
|
| 1747 |
float32 float32_add( float32 a, float32 b STATUS_PARAM ) |
| 1748 |
{
|
| 1749 |
flag aSign, bSign; |
| 1750 |
|
| 1751 |
aSign = extractFloat32Sign( a ); |
| 1752 |
bSign = extractFloat32Sign( b ); |
| 1753 |
if ( aSign == bSign ) {
|
| 1754 |
return addFloat32Sigs( a, b, aSign STATUS_VAR);
|
| 1755 |
} |
| 1756 |
else {
|
| 1757 |
return subFloat32Sigs( a, b, aSign STATUS_VAR );
|
| 1758 |
} |
| 1759 |
|
| 1760 |
} |
| 1761 |
|
| 1762 |
/*----------------------------------------------------------------------------
|
| 1763 |
| Returns the result of subtracting the single-precision floating-point values
|
| 1764 |
| `a' and `b'. The operation is performed according to the IEC/IEEE Standard
|
| 1765 |
| for Binary Floating-Point Arithmetic.
|
| 1766 |
*----------------------------------------------------------------------------*/
|
| 1767 |
|
| 1768 |
float32 float32_sub( float32 a, float32 b STATUS_PARAM ) |
| 1769 |
{
|
| 1770 |
flag aSign, bSign; |
| 1771 |
|
| 1772 |
aSign = extractFloat32Sign( a ); |
| 1773 |
bSign = extractFloat32Sign( b ); |
| 1774 |
if ( aSign == bSign ) {
|
| 1775 |
return subFloat32Sigs( a, b, aSign STATUS_VAR );
|
| 1776 |
} |
| 1777 |
else {
|
| 1778 |
return addFloat32Sigs( a, b, aSign STATUS_VAR );
|
| 1779 |
} |
| 1780 |
|
| 1781 |
} |
| 1782 |
|
| 1783 |
/*----------------------------------------------------------------------------
|
| 1784 |
| Returns the result of multiplying the single-precision floating-point values
|
| 1785 |
| `a' and `b'. The operation is performed according to the IEC/IEEE Standard
|
| 1786 |
| for Binary Floating-Point Arithmetic.
|
| 1787 |
*----------------------------------------------------------------------------*/
|
| 1788 |
|
| 1789 |
float32 float32_mul( float32 a, float32 b STATUS_PARAM ) |
| 1790 |
{
|
| 1791 |
flag aSign, bSign, zSign; |
| 1792 |
int16 aExp, bExp, zExp; |
| 1793 |
bits32 aSig, bSig; |
| 1794 |
bits64 zSig64; |
| 1795 |
bits32 zSig; |
| 1796 |
|
| 1797 |
aSig = extractFloat32Frac( a ); |
| 1798 |
aExp = extractFloat32Exp( a ); |
| 1799 |
aSign = extractFloat32Sign( a ); |
| 1800 |
bSig = extractFloat32Frac( b ); |
| 1801 |
bExp = extractFloat32Exp( b ); |
| 1802 |
bSign = extractFloat32Sign( b ); |
| 1803 |
zSign = aSign ^ bSign; |
| 1804 |
if ( aExp == 0xFF ) { |
| 1805 |
if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { |
| 1806 |
return propagateFloat32NaN( a, b STATUS_VAR );
|
| 1807 |
} |
| 1808 |
if ( ( bExp | bSig ) == 0 ) { |
| 1809 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1810 |
return float32_default_nan;
|
| 1811 |
} |
| 1812 |
return packFloat32( zSign, 0xFF, 0 ); |
| 1813 |
} |
| 1814 |
if ( bExp == 0xFF ) { |
| 1815 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1816 |
if ( ( aExp | aSig ) == 0 ) { |
| 1817 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1818 |
return float32_default_nan;
|
| 1819 |
} |
| 1820 |
return packFloat32( zSign, 0xFF, 0 ); |
| 1821 |
} |
| 1822 |
if ( aExp == 0 ) { |
| 1823 |
if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); |
| 1824 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1825 |
} |
| 1826 |
if ( bExp == 0 ) { |
| 1827 |
if ( bSig == 0 ) return packFloat32( zSign, 0, 0 ); |
| 1828 |
normalizeFloat32Subnormal( bSig, &bExp, &bSig ); |
| 1829 |
} |
| 1830 |
zExp = aExp + bExp - 0x7F;
|
| 1831 |
aSig = ( aSig | 0x00800000 )<<7; |
| 1832 |
bSig = ( bSig | 0x00800000 )<<8; |
| 1833 |
shift64RightJamming( ( (bits64) aSig ) * bSig, 32, &zSig64 );
|
| 1834 |
zSig = zSig64; |
| 1835 |
if ( 0 <= (sbits32) ( zSig<<1 ) ) { |
| 1836 |
zSig <<= 1;
|
| 1837 |
--zExp; |
| 1838 |
} |
| 1839 |
return roundAndPackFloat32( zSign, zExp, zSig STATUS_VAR );
|
| 1840 |
|
| 1841 |
} |
| 1842 |
|
| 1843 |
/*----------------------------------------------------------------------------
|
| 1844 |
| Returns the result of dividing the single-precision floating-point value `a'
|
| 1845 |
| by the corresponding value `b'. The operation is performed according to the
|
| 1846 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1847 |
*----------------------------------------------------------------------------*/
|
| 1848 |
|
| 1849 |
float32 float32_div( float32 a, float32 b STATUS_PARAM ) |
| 1850 |
{
|
| 1851 |
flag aSign, bSign, zSign; |
| 1852 |
int16 aExp, bExp, zExp; |
| 1853 |
bits32 aSig, bSig, zSig; |
| 1854 |
|
| 1855 |
aSig = extractFloat32Frac( a ); |
| 1856 |
aExp = extractFloat32Exp( a ); |
| 1857 |
aSign = extractFloat32Sign( a ); |
| 1858 |
bSig = extractFloat32Frac( b ); |
| 1859 |
bExp = extractFloat32Exp( b ); |
| 1860 |
bSign = extractFloat32Sign( b ); |
| 1861 |
zSign = aSign ^ bSign; |
| 1862 |
if ( aExp == 0xFF ) { |
| 1863 |
if ( aSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1864 |
if ( bExp == 0xFF ) { |
| 1865 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1866 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1867 |
return float32_default_nan;
|
| 1868 |
} |
| 1869 |
return packFloat32( zSign, 0xFF, 0 ); |
| 1870 |
} |
| 1871 |
if ( bExp == 0xFF ) { |
| 1872 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1873 |
return packFloat32( zSign, 0, 0 ); |
| 1874 |
} |
| 1875 |
if ( bExp == 0 ) { |
| 1876 |
if ( bSig == 0 ) { |
| 1877 |
if ( ( aExp | aSig ) == 0 ) { |
| 1878 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1879 |
return float32_default_nan;
|
| 1880 |
} |
| 1881 |
float_raise( float_flag_divbyzero STATUS_VAR); |
| 1882 |
return packFloat32( zSign, 0xFF, 0 ); |
| 1883 |
} |
| 1884 |
normalizeFloat32Subnormal( bSig, &bExp, &bSig ); |
| 1885 |
} |
| 1886 |
if ( aExp == 0 ) { |
| 1887 |
if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); |
| 1888 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1889 |
} |
| 1890 |
zExp = aExp - bExp + 0x7D;
|
| 1891 |
aSig = ( aSig | 0x00800000 )<<7; |
| 1892 |
bSig = ( bSig | 0x00800000 )<<8; |
| 1893 |
if ( bSig <= ( aSig + aSig ) ) {
|
| 1894 |
aSig >>= 1;
|
| 1895 |
++zExp; |
| 1896 |
} |
| 1897 |
zSig = ( ( (bits64) aSig )<<32 ) / bSig;
|
| 1898 |
if ( ( zSig & 0x3F ) == 0 ) { |
| 1899 |
zSig |= ( (bits64) bSig * zSig != ( (bits64) aSig )<<32 );
|
| 1900 |
} |
| 1901 |
return roundAndPackFloat32( zSign, zExp, zSig STATUS_VAR );
|
| 1902 |
|
| 1903 |
} |
| 1904 |
|
| 1905 |
/*----------------------------------------------------------------------------
|
| 1906 |
| Returns the remainder of the single-precision floating-point value `a'
|
| 1907 |
| with respect to the corresponding value `b'. The operation is performed
|
| 1908 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 1909 |
*----------------------------------------------------------------------------*/
|
| 1910 |
|
| 1911 |
float32 float32_rem( float32 a, float32 b STATUS_PARAM ) |
| 1912 |
{
|
| 1913 |
flag aSign, zSign; |
| 1914 |
int16 aExp, bExp, expDiff; |
| 1915 |
bits32 aSig, bSig; |
| 1916 |
bits32 q; |
| 1917 |
bits64 aSig64, bSig64, q64; |
| 1918 |
bits32 alternateASig; |
| 1919 |
sbits32 sigMean; |
| 1920 |
|
| 1921 |
aSig = extractFloat32Frac( a ); |
| 1922 |
aExp = extractFloat32Exp( a ); |
| 1923 |
aSign = extractFloat32Sign( a ); |
| 1924 |
bSig = extractFloat32Frac( b ); |
| 1925 |
bExp = extractFloat32Exp( b ); |
| 1926 |
if ( aExp == 0xFF ) { |
| 1927 |
if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { |
| 1928 |
return propagateFloat32NaN( a, b STATUS_VAR );
|
| 1929 |
} |
| 1930 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1931 |
return float32_default_nan;
|
| 1932 |
} |
| 1933 |
if ( bExp == 0xFF ) { |
| 1934 |
if ( bSig ) return propagateFloat32NaN( a, b STATUS_VAR ); |
| 1935 |
return a;
|
| 1936 |
} |
| 1937 |
if ( bExp == 0 ) { |
| 1938 |
if ( bSig == 0 ) { |
| 1939 |
float_raise( float_flag_invalid STATUS_VAR); |
| 1940 |
return float32_default_nan;
|
| 1941 |
} |
| 1942 |
normalizeFloat32Subnormal( bSig, &bExp, &bSig ); |
| 1943 |
} |
| 1944 |
if ( aExp == 0 ) { |
| 1945 |
if ( aSig == 0 ) return a; |
| 1946 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 1947 |
} |
| 1948 |
expDiff = aExp - bExp; |
| 1949 |
aSig |= 0x00800000;
|
| 1950 |
bSig |= 0x00800000;
|
| 1951 |
if ( expDiff < 32 ) { |
| 1952 |
aSig <<= 8;
|
| 1953 |
bSig <<= 8;
|
| 1954 |
if ( expDiff < 0 ) { |
| 1955 |
if ( expDiff < -1 ) return a; |
| 1956 |
aSig >>= 1;
|
| 1957 |
} |
| 1958 |
q = ( bSig <= aSig ); |
| 1959 |
if ( q ) aSig -= bSig;
|
| 1960 |
if ( 0 < expDiff ) { |
| 1961 |
q = ( ( (bits64) aSig )<<32 ) / bSig;
|
| 1962 |
q >>= 32 - expDiff;
|
| 1963 |
bSig >>= 2;
|
| 1964 |
aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; |
| 1965 |
} |
| 1966 |
else {
|
| 1967 |
aSig >>= 2;
|
| 1968 |
bSig >>= 2;
|
| 1969 |
} |
| 1970 |
} |
| 1971 |
else {
|
| 1972 |
if ( bSig <= aSig ) aSig -= bSig;
|
| 1973 |
aSig64 = ( (bits64) aSig )<<40;
|
| 1974 |
bSig64 = ( (bits64) bSig )<<40;
|
| 1975 |
expDiff -= 64;
|
| 1976 |
while ( 0 < expDiff ) { |
| 1977 |
q64 = estimateDiv128To64( aSig64, 0, bSig64 );
|
| 1978 |
q64 = ( 2 < q64 ) ? q64 - 2 : 0; |
| 1979 |
aSig64 = - ( ( bSig * q64 )<<38 );
|
| 1980 |
expDiff -= 62;
|
| 1981 |
} |
| 1982 |
expDiff += 64;
|
| 1983 |
q64 = estimateDiv128To64( aSig64, 0, bSig64 );
|
| 1984 |
q64 = ( 2 < q64 ) ? q64 - 2 : 0; |
| 1985 |
q = q64>>( 64 - expDiff );
|
| 1986 |
bSig <<= 6;
|
| 1987 |
aSig = ( ( aSig64>>33 )<<( expDiff - 1 ) ) - bSig * q; |
| 1988 |
} |
| 1989 |
do {
|
| 1990 |
alternateASig = aSig; |
| 1991 |
++q; |
| 1992 |
aSig -= bSig; |
| 1993 |
} while ( 0 <= (sbits32) aSig ); |
| 1994 |
sigMean = aSig + alternateASig; |
| 1995 |
if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { |
| 1996 |
aSig = alternateASig; |
| 1997 |
} |
| 1998 |
zSign = ( (sbits32) aSig < 0 );
|
| 1999 |
if ( zSign ) aSig = - aSig;
|
| 2000 |
return normalizeRoundAndPackFloat32( aSign ^ zSign, bExp, aSig STATUS_VAR );
|
| 2001 |
|
| 2002 |
} |
| 2003 |
|
| 2004 |
/*----------------------------------------------------------------------------
|
| 2005 |
| Returns the square root of the single-precision floating-point value `a'.
|
| 2006 |
| The operation is performed according to the IEC/IEEE Standard for Binary
|
| 2007 |
| Floating-Point Arithmetic.
|
| 2008 |
*----------------------------------------------------------------------------*/
|
| 2009 |
|
| 2010 |
float32 float32_sqrt( float32 a STATUS_PARAM ) |
| 2011 |
{
|
| 2012 |
flag aSign; |
| 2013 |
int16 aExp, zExp; |
| 2014 |
bits32 aSig, zSig; |
| 2015 |
bits64 rem, term; |
| 2016 |
|
| 2017 |
aSig = extractFloat32Frac( a ); |
| 2018 |
aExp = extractFloat32Exp( a ); |
| 2019 |
aSign = extractFloat32Sign( a ); |
| 2020 |
if ( aExp == 0xFF ) { |
| 2021 |
if ( aSig ) return propagateFloat32NaN( a, float32_zero STATUS_VAR ); |
| 2022 |
if ( ! aSign ) return a; |
| 2023 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2024 |
return float32_default_nan;
|
| 2025 |
} |
| 2026 |
if ( aSign ) {
|
| 2027 |
if ( ( aExp | aSig ) == 0 ) return a; |
| 2028 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2029 |
return float32_default_nan;
|
| 2030 |
} |
| 2031 |
if ( aExp == 0 ) { |
| 2032 |
if ( aSig == 0 ) return float32_zero; |
| 2033 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 2034 |
} |
| 2035 |
zExp = ( ( aExp - 0x7F )>>1 ) + 0x7E; |
| 2036 |
aSig = ( aSig | 0x00800000 )<<8; |
| 2037 |
zSig = estimateSqrt32( aExp, aSig ) + 2;
|
| 2038 |
if ( ( zSig & 0x7F ) <= 5 ) { |
| 2039 |
if ( zSig < 2 ) { |
| 2040 |
zSig = 0x7FFFFFFF;
|
| 2041 |
goto roundAndPack;
|
| 2042 |
} |
| 2043 |
aSig >>= aExp & 1;
|
| 2044 |
term = ( (bits64) zSig ) * zSig; |
| 2045 |
rem = ( ( (bits64) aSig )<<32 ) - term;
|
| 2046 |
while ( (sbits64) rem < 0 ) { |
| 2047 |
--zSig; |
| 2048 |
rem += ( ( (bits64) zSig )<<1 ) | 1; |
| 2049 |
} |
| 2050 |
zSig |= ( rem != 0 );
|
| 2051 |
} |
| 2052 |
shift32RightJamming( zSig, 1, &zSig );
|
| 2053 |
roundAndPack:
|
| 2054 |
return roundAndPackFloat32( 0, zExp, zSig STATUS_VAR ); |
| 2055 |
|
| 2056 |
} |
| 2057 |
|
| 2058 |
/*----------------------------------------------------------------------------
|
| 2059 |
| Returns the binary log of the single-precision floating-point value `a'.
|
| 2060 |
| The operation is performed according to the IEC/IEEE Standard for Binary
|
| 2061 |
| Floating-Point Arithmetic.
|
| 2062 |
*----------------------------------------------------------------------------*/
|
| 2063 |
float32 float32_log2( float32 a STATUS_PARAM ) |
| 2064 |
{
|
| 2065 |
flag aSign, zSign; |
| 2066 |
int16 aExp; |
| 2067 |
bits32 aSig, zSig, i; |
| 2068 |
|
| 2069 |
aSig = extractFloat32Frac( a ); |
| 2070 |
aExp = extractFloat32Exp( a ); |
| 2071 |
aSign = extractFloat32Sign( a ); |
| 2072 |
|
| 2073 |
if ( aExp == 0 ) { |
| 2074 |
if ( aSig == 0 ) return packFloat32( 1, 0xFF, 0 ); |
| 2075 |
normalizeFloat32Subnormal( aSig, &aExp, &aSig ); |
| 2076 |
} |
| 2077 |
if ( aSign ) {
|
| 2078 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2079 |
return float32_default_nan;
|
| 2080 |
} |
| 2081 |
if ( aExp == 0xFF ) { |
| 2082 |
if ( aSig ) return propagateFloat32NaN( a, float32_zero STATUS_VAR ); |
| 2083 |
return a;
|
| 2084 |
} |
| 2085 |
|
| 2086 |
aExp -= 0x7F;
|
| 2087 |
aSig |= 0x00800000;
|
| 2088 |
zSign = aExp < 0;
|
| 2089 |
zSig = aExp << 23;
|
| 2090 |
|
| 2091 |
for (i = 1 << 22; i > 0; i >>= 1) { |
| 2092 |
aSig = ( (bits64)aSig * aSig ) >> 23;
|
| 2093 |
if ( aSig & 0x01000000 ) { |
| 2094 |
aSig >>= 1;
|
| 2095 |
zSig |= i; |
| 2096 |
} |
| 2097 |
} |
| 2098 |
|
| 2099 |
if ( zSign )
|
| 2100 |
zSig = -zSig; |
| 2101 |
|
| 2102 |
return normalizeRoundAndPackFloat32( zSign, 0x85, zSig STATUS_VAR ); |
| 2103 |
} |
| 2104 |
|
| 2105 |
/*----------------------------------------------------------------------------
|
| 2106 |
| Returns 1 if the single-precision floating-point value `a' is equal to
|
| 2107 |
| the corresponding value `b', and 0 otherwise. The comparison is performed
|
| 2108 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 2109 |
*----------------------------------------------------------------------------*/
|
| 2110 |
|
| 2111 |
int float32_eq( float32 a, float32 b STATUS_PARAM )
|
| 2112 |
{
|
| 2113 |
|
| 2114 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2115 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2116 |
) {
|
| 2117 |
if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) {
|
| 2118 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2119 |
} |
| 2120 |
return 0; |
| 2121 |
} |
| 2122 |
return ( float32_val(a) == float32_val(b) ) ||
|
| 2123 |
( (bits32) ( ( float32_val(a) | float32_val(b) )<<1 ) == 0 ); |
| 2124 |
|
| 2125 |
} |
| 2126 |
|
| 2127 |
/*----------------------------------------------------------------------------
|
| 2128 |
| Returns 1 if the single-precision floating-point value `a' is less than
|
| 2129 |
| or equal to the corresponding value `b', and 0 otherwise. The comparison
|
| 2130 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2131 |
| Arithmetic.
|
| 2132 |
*----------------------------------------------------------------------------*/
|
| 2133 |
|
| 2134 |
int float32_le( float32 a, float32 b STATUS_PARAM )
|
| 2135 |
{
|
| 2136 |
flag aSign, bSign; |
| 2137 |
bits32 av, bv; |
| 2138 |
|
| 2139 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2140 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2141 |
) {
|
| 2142 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2143 |
return 0; |
| 2144 |
} |
| 2145 |
aSign = extractFloat32Sign( a ); |
| 2146 |
bSign = extractFloat32Sign( b ); |
| 2147 |
av = float32_val(a); |
| 2148 |
bv = float32_val(b); |
| 2149 |
if ( aSign != bSign ) return aSign || ( (bits32) ( ( av | bv )<<1 ) == 0 ); |
| 2150 |
return ( av == bv ) || ( aSign ^ ( av < bv ) );
|
| 2151 |
|
| 2152 |
} |
| 2153 |
|
| 2154 |
/*----------------------------------------------------------------------------
|
| 2155 |
| Returns 1 if the single-precision floating-point value `a' is less than
|
| 2156 |
| the corresponding value `b', and 0 otherwise. The comparison is performed
|
| 2157 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 2158 |
*----------------------------------------------------------------------------*/
|
| 2159 |
|
| 2160 |
int float32_lt( float32 a, float32 b STATUS_PARAM )
|
| 2161 |
{
|
| 2162 |
flag aSign, bSign; |
| 2163 |
bits32 av, bv; |
| 2164 |
|
| 2165 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2166 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2167 |
) {
|
| 2168 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2169 |
return 0; |
| 2170 |
} |
| 2171 |
aSign = extractFloat32Sign( a ); |
| 2172 |
bSign = extractFloat32Sign( b ); |
| 2173 |
av = float32_val(a); |
| 2174 |
bv = float32_val(b); |
| 2175 |
if ( aSign != bSign ) return aSign && ( (bits32) ( ( av | bv )<<1 ) != 0 ); |
| 2176 |
return ( av != bv ) && ( aSign ^ ( av < bv ) );
|
| 2177 |
|
| 2178 |
} |
| 2179 |
|
| 2180 |
/*----------------------------------------------------------------------------
|
| 2181 |
| Returns 1 if the single-precision floating-point value `a' is equal to
|
| 2182 |
| the corresponding value `b', and 0 otherwise. The invalid exception is
|
| 2183 |
| raised if either operand is a NaN. Otherwise, the comparison is performed
|
| 2184 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 2185 |
*----------------------------------------------------------------------------*/
|
| 2186 |
|
| 2187 |
int float32_eq_signaling( float32 a, float32 b STATUS_PARAM )
|
| 2188 |
{
|
| 2189 |
bits32 av, bv; |
| 2190 |
|
| 2191 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2192 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2193 |
) {
|
| 2194 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2195 |
return 0; |
| 2196 |
} |
| 2197 |
av = float32_val(a); |
| 2198 |
bv = float32_val(b); |
| 2199 |
return ( av == bv ) || ( (bits32) ( ( av | bv )<<1 ) == 0 ); |
| 2200 |
|
| 2201 |
} |
| 2202 |
|
| 2203 |
/*----------------------------------------------------------------------------
|
| 2204 |
| Returns 1 if the single-precision floating-point value `a' is less than or
|
| 2205 |
| equal to the corresponding value `b', and 0 otherwise. Quiet NaNs do not
|
| 2206 |
| cause an exception. Otherwise, the comparison is performed according to the
|
| 2207 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 2208 |
*----------------------------------------------------------------------------*/
|
| 2209 |
|
| 2210 |
int float32_le_quiet( float32 a, float32 b STATUS_PARAM )
|
| 2211 |
{
|
| 2212 |
flag aSign, bSign; |
| 2213 |
bits32 av, bv; |
| 2214 |
|
| 2215 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2216 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2217 |
) {
|
| 2218 |
if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) {
|
| 2219 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2220 |
} |
| 2221 |
return 0; |
| 2222 |
} |
| 2223 |
aSign = extractFloat32Sign( a ); |
| 2224 |
bSign = extractFloat32Sign( b ); |
| 2225 |
av = float32_val(a); |
| 2226 |
bv = float32_val(b); |
| 2227 |
if ( aSign != bSign ) return aSign || ( (bits32) ( ( av | bv )<<1 ) == 0 ); |
| 2228 |
return ( av == bv ) || ( aSign ^ ( av < bv ) );
|
| 2229 |
|
| 2230 |
} |
| 2231 |
|
| 2232 |
/*----------------------------------------------------------------------------
|
| 2233 |
| Returns 1 if the single-precision floating-point value `a' is less than
|
| 2234 |
| the corresponding value `b', and 0 otherwise. Quiet NaNs do not cause an
|
| 2235 |
| exception. Otherwise, the comparison is performed according to the IEC/IEEE
|
| 2236 |
| Standard for Binary Floating-Point Arithmetic.
|
| 2237 |
*----------------------------------------------------------------------------*/
|
| 2238 |
|
| 2239 |
int float32_lt_quiet( float32 a, float32 b STATUS_PARAM )
|
| 2240 |
{
|
| 2241 |
flag aSign, bSign; |
| 2242 |
bits32 av, bv; |
| 2243 |
|
| 2244 |
if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) |
| 2245 |
|| ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) )
|
| 2246 |
) {
|
| 2247 |
if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) {
|
| 2248 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2249 |
} |
| 2250 |
return 0; |
| 2251 |
} |
| 2252 |
aSign = extractFloat32Sign( a ); |
| 2253 |
bSign = extractFloat32Sign( b ); |
| 2254 |
av = float32_val(a); |
| 2255 |
bv = float32_val(b); |
| 2256 |
if ( aSign != bSign ) return aSign && ( (bits32) ( ( av | bv )<<1 ) != 0 ); |
| 2257 |
return ( av != bv ) && ( aSign ^ ( av < bv ) );
|
| 2258 |
|
| 2259 |
} |
| 2260 |
|
| 2261 |
/*----------------------------------------------------------------------------
|
| 2262 |
| Returns the result of converting the double-precision floating-point value
|
| 2263 |
| `a' to the 32-bit two's complement integer format. The conversion is
|
| 2264 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2265 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 2266 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 2267 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 2268 |
| largest integer with the same sign as `a' is returned.
|
| 2269 |
*----------------------------------------------------------------------------*/
|
| 2270 |
|
| 2271 |
int32 float64_to_int32( float64 a STATUS_PARAM ) |
| 2272 |
{
|
| 2273 |
flag aSign; |
| 2274 |
int16 aExp, shiftCount; |
| 2275 |
bits64 aSig; |
| 2276 |
|
| 2277 |
aSig = extractFloat64Frac( a ); |
| 2278 |
aExp = extractFloat64Exp( a ); |
| 2279 |
aSign = extractFloat64Sign( a ); |
| 2280 |
if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; |
| 2281 |
if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); |
| 2282 |
shiftCount = 0x42C - aExp;
|
| 2283 |
if ( 0 < shiftCount ) shift64RightJamming( aSig, shiftCount, &aSig ); |
| 2284 |
return roundAndPackInt32( aSign, aSig STATUS_VAR );
|
| 2285 |
|
| 2286 |
} |
| 2287 |
|
| 2288 |
/*----------------------------------------------------------------------------
|
| 2289 |
| Returns the result of converting the double-precision floating-point value
|
| 2290 |
| `a' to the 32-bit two's complement integer format. The conversion is
|
| 2291 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2292 |
| Arithmetic, except that the conversion is always rounded toward zero.
|
| 2293 |
| If `a' is a NaN, the largest positive integer is returned. Otherwise, if
|
| 2294 |
| the conversion overflows, the largest integer with the same sign as `a' is
|
| 2295 |
| returned.
|
| 2296 |
*----------------------------------------------------------------------------*/
|
| 2297 |
|
| 2298 |
int32 float64_to_int32_round_to_zero( float64 a STATUS_PARAM ) |
| 2299 |
{
|
| 2300 |
flag aSign; |
| 2301 |
int16 aExp, shiftCount; |
| 2302 |
bits64 aSig, savedASig; |
| 2303 |
int32 z; |
| 2304 |
|
| 2305 |
aSig = extractFloat64Frac( a ); |
| 2306 |
aExp = extractFloat64Exp( a ); |
| 2307 |
aSign = extractFloat64Sign( a ); |
| 2308 |
if ( 0x41E < aExp ) { |
| 2309 |
if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; |
| 2310 |
goto invalid;
|
| 2311 |
} |
| 2312 |
else if ( aExp < 0x3FF ) { |
| 2313 |
if ( aExp || aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 2314 |
return 0; |
| 2315 |
} |
| 2316 |
aSig |= LIT64( 0x0010000000000000 );
|
| 2317 |
shiftCount = 0x433 - aExp;
|
| 2318 |
savedASig = aSig; |
| 2319 |
aSig >>= shiftCount; |
| 2320 |
z = aSig; |
| 2321 |
if ( aSign ) z = - z;
|
| 2322 |
if ( ( z < 0 ) ^ aSign ) { |
| 2323 |
invalid:
|
| 2324 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2325 |
return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; |
| 2326 |
} |
| 2327 |
if ( ( aSig<<shiftCount ) != savedASig ) {
|
| 2328 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 2329 |
} |
| 2330 |
return z;
|
| 2331 |
|
| 2332 |
} |
| 2333 |
|
| 2334 |
/*----------------------------------------------------------------------------
|
| 2335 |
| Returns the result of converting the double-precision floating-point value
|
| 2336 |
| `a' to the 64-bit two's complement integer format. The conversion is
|
| 2337 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2338 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 2339 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 2340 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 2341 |
| largest integer with the same sign as `a' is returned.
|
| 2342 |
*----------------------------------------------------------------------------*/
|
| 2343 |
|
| 2344 |
int64 float64_to_int64( float64 a STATUS_PARAM ) |
| 2345 |
{
|
| 2346 |
flag aSign; |
| 2347 |
int16 aExp, shiftCount; |
| 2348 |
bits64 aSig, aSigExtra; |
| 2349 |
|
| 2350 |
aSig = extractFloat64Frac( a ); |
| 2351 |
aExp = extractFloat64Exp( a ); |
| 2352 |
aSign = extractFloat64Sign( a ); |
| 2353 |
if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); |
| 2354 |
shiftCount = 0x433 - aExp;
|
| 2355 |
if ( shiftCount <= 0 ) { |
| 2356 |
if ( 0x43E < aExp ) { |
| 2357 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2358 |
if ( ! aSign
|
| 2359 |
|| ( ( aExp == 0x7FF )
|
| 2360 |
&& ( aSig != LIT64( 0x0010000000000000 ) ) )
|
| 2361 |
) {
|
| 2362 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 2363 |
} |
| 2364 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 2365 |
} |
| 2366 |
aSigExtra = 0;
|
| 2367 |
aSig <<= - shiftCount; |
| 2368 |
} |
| 2369 |
else {
|
| 2370 |
shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra );
|
| 2371 |
} |
| 2372 |
return roundAndPackInt64( aSign, aSig, aSigExtra STATUS_VAR );
|
| 2373 |
|
| 2374 |
} |
| 2375 |
|
| 2376 |
/*----------------------------------------------------------------------------
|
| 2377 |
| Returns the result of converting the double-precision floating-point value
|
| 2378 |
| `a' to the 64-bit two's complement integer format. The conversion is
|
| 2379 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2380 |
| Arithmetic, except that the conversion is always rounded toward zero.
|
| 2381 |
| If `a' is a NaN, the largest positive integer is returned. Otherwise, if
|
| 2382 |
| the conversion overflows, the largest integer with the same sign as `a' is
|
| 2383 |
| returned.
|
| 2384 |
*----------------------------------------------------------------------------*/
|
| 2385 |
|
| 2386 |
int64 float64_to_int64_round_to_zero( float64 a STATUS_PARAM ) |
| 2387 |
{
|
| 2388 |
flag aSign; |
| 2389 |
int16 aExp, shiftCount; |
| 2390 |
bits64 aSig; |
| 2391 |
int64 z; |
| 2392 |
|
| 2393 |
aSig = extractFloat64Frac( a ); |
| 2394 |
aExp = extractFloat64Exp( a ); |
| 2395 |
aSign = extractFloat64Sign( a ); |
| 2396 |
if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); |
| 2397 |
shiftCount = aExp - 0x433;
|
| 2398 |
if ( 0 <= shiftCount ) { |
| 2399 |
if ( 0x43E <= aExp ) { |
| 2400 |
if ( float64_val(a) != LIT64( 0xC3E0000000000000 ) ) { |
| 2401 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2402 |
if ( ! aSign
|
| 2403 |
|| ( ( aExp == 0x7FF )
|
| 2404 |
&& ( aSig != LIT64( 0x0010000000000000 ) ) )
|
| 2405 |
) {
|
| 2406 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 2407 |
} |
| 2408 |
} |
| 2409 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 2410 |
} |
| 2411 |
z = aSig<<shiftCount; |
| 2412 |
} |
| 2413 |
else {
|
| 2414 |
if ( aExp < 0x3FE ) { |
| 2415 |
if ( aExp | aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 2416 |
return 0; |
| 2417 |
} |
| 2418 |
z = aSig>>( - shiftCount ); |
| 2419 |
if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { |
| 2420 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 2421 |
} |
| 2422 |
} |
| 2423 |
if ( aSign ) z = - z;
|
| 2424 |
return z;
|
| 2425 |
|
| 2426 |
} |
| 2427 |
|
| 2428 |
/*----------------------------------------------------------------------------
|
| 2429 |
| Returns the result of converting the double-precision floating-point value
|
| 2430 |
| `a' to the single-precision floating-point format. The conversion is
|
| 2431 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2432 |
| Arithmetic.
|
| 2433 |
*----------------------------------------------------------------------------*/
|
| 2434 |
|
| 2435 |
float32 float64_to_float32( float64 a STATUS_PARAM ) |
| 2436 |
{
|
| 2437 |
flag aSign; |
| 2438 |
int16 aExp; |
| 2439 |
bits64 aSig; |
| 2440 |
bits32 zSig; |
| 2441 |
|
| 2442 |
aSig = extractFloat64Frac( a ); |
| 2443 |
aExp = extractFloat64Exp( a ); |
| 2444 |
aSign = extractFloat64Sign( a ); |
| 2445 |
if ( aExp == 0x7FF ) { |
| 2446 |
if ( aSig ) return commonNaNToFloat32( float64ToCommonNaN( a STATUS_VAR ) ); |
| 2447 |
return packFloat32( aSign, 0xFF, 0 ); |
| 2448 |
} |
| 2449 |
shift64RightJamming( aSig, 22, &aSig );
|
| 2450 |
zSig = aSig; |
| 2451 |
if ( aExp || zSig ) {
|
| 2452 |
zSig |= 0x40000000;
|
| 2453 |
aExp -= 0x381;
|
| 2454 |
} |
| 2455 |
return roundAndPackFloat32( aSign, aExp, zSig STATUS_VAR );
|
| 2456 |
|
| 2457 |
} |
| 2458 |
|
| 2459 |
|
| 2460 |
/*----------------------------------------------------------------------------
|
| 2461 |
| Packs the sign `zSign', exponent `zExp', and significand `zSig' into a
|
| 2462 |
| half-precision floating-point value, returning the result. After being
|
| 2463 |
| shifted into the proper positions, the three fields are simply added
|
| 2464 |
| together to form the result. This means that any integer portion of `zSig'
|
| 2465 |
| will be added into the exponent. Since a properly normalized significand
|
| 2466 |
| will have an integer portion equal to 1, the `zExp' input should be 1 less
|
| 2467 |
| than the desired result exponent whenever `zSig' is a complete, normalized
|
| 2468 |
| significand.
|
| 2469 |
*----------------------------------------------------------------------------*/
|
| 2470 |
static bits16 packFloat16(flag zSign, int16 zExp, bits16 zSig)
|
| 2471 |
{
|
| 2472 |
return (((bits32)zSign) << 15) + (((bits32)zExp) << 10) + zSig; |
| 2473 |
} |
| 2474 |
|
| 2475 |
/* Half precision floats come in two formats: standard IEEE and "ARM" format.
|
| 2476 |
The latter gains extra exponent range by omitting the NaN/Inf encodings. */
|
| 2477 |
|
| 2478 |
float32 float16_to_float32( bits16 a, flag ieee STATUS_PARAM ) |
| 2479 |
{
|
| 2480 |
flag aSign; |
| 2481 |
int16 aExp; |
| 2482 |
bits32 aSig; |
| 2483 |
|
| 2484 |
aSign = a >> 15;
|
| 2485 |
aExp = (a >> 10) & 0x1f; |
| 2486 |
aSig = a & 0x3ff;
|
| 2487 |
|
| 2488 |
if (aExp == 0x1f && ieee) { |
| 2489 |
if (aSig) {
|
| 2490 |
/* Make sure correct exceptions are raised. */
|
| 2491 |
float32ToCommonNaN(a STATUS_VAR); |
| 2492 |
aSig |= 0x200;
|
| 2493 |
} |
| 2494 |
return packFloat32(aSign, 0xff, aSig << 13); |
| 2495 |
} |
| 2496 |
if (aExp == 0) { |
| 2497 |
int8 shiftCount; |
| 2498 |
|
| 2499 |
if (aSig == 0) { |
| 2500 |
return packFloat32(aSign, 0, 0); |
| 2501 |
} |
| 2502 |
|
| 2503 |
shiftCount = countLeadingZeros32( aSig ) - 21;
|
| 2504 |
aSig = aSig << shiftCount; |
| 2505 |
aExp = -shiftCount; |
| 2506 |
} |
| 2507 |
return packFloat32( aSign, aExp + 0x70, aSig << 13); |
| 2508 |
} |
| 2509 |
|
| 2510 |
bits16 float32_to_float16( float32 a, flag ieee STATUS_PARAM) |
| 2511 |
{
|
| 2512 |
flag aSign; |
| 2513 |
int16 aExp; |
| 2514 |
bits32 aSig; |
| 2515 |
bits32 mask; |
| 2516 |
bits32 increment; |
| 2517 |
int8 roundingMode; |
| 2518 |
|
| 2519 |
aSig = extractFloat32Frac( a ); |
| 2520 |
aExp = extractFloat32Exp( a ); |
| 2521 |
aSign = extractFloat32Sign( a ); |
| 2522 |
if ( aExp == 0xFF ) { |
| 2523 |
if (aSig) {
|
| 2524 |
/* Make sure correct exceptions are raised. */
|
| 2525 |
float32ToCommonNaN(a STATUS_VAR); |
| 2526 |
aSig |= 0x00400000;
|
| 2527 |
} |
| 2528 |
return packFloat16(aSign, 0x1f, aSig >> 13); |
| 2529 |
} |
| 2530 |
if (aExp == 0 && aSign == 0) { |
| 2531 |
return packFloat16(aSign, 0, 0); |
| 2532 |
} |
| 2533 |
/* Decimal point between bits 22 and 23. */
|
| 2534 |
aSig |= 0x00800000;
|
| 2535 |
aExp -= 0x7f;
|
| 2536 |
if (aExp < -14) { |
| 2537 |
mask = 0x007fffff;
|
| 2538 |
if (aExp < -24) { |
| 2539 |
aExp = -25;
|
| 2540 |
} else {
|
| 2541 |
mask >>= 24 + aExp;
|
| 2542 |
} |
| 2543 |
} else {
|
| 2544 |
mask = 0x00001fff;
|
| 2545 |
} |
| 2546 |
if (aSig & mask) {
|
| 2547 |
float_raise( float_flag_underflow STATUS_VAR ); |
| 2548 |
roundingMode = STATUS(float_rounding_mode); |
| 2549 |
switch (roundingMode) {
|
| 2550 |
case float_round_nearest_even:
|
| 2551 |
increment = (mask + 1) >> 1; |
| 2552 |
if ((aSig & mask) == increment) {
|
| 2553 |
increment = aSig & (increment << 1);
|
| 2554 |
} |
| 2555 |
break;
|
| 2556 |
case float_round_up:
|
| 2557 |
increment = aSign ? 0 : mask;
|
| 2558 |
break;
|
| 2559 |
case float_round_down:
|
| 2560 |
increment = aSign ? mask : 0;
|
| 2561 |
break;
|
| 2562 |
default: /* round_to_zero */ |
| 2563 |
increment = 0;
|
| 2564 |
break;
|
| 2565 |
} |
| 2566 |
aSig += increment; |
| 2567 |
if (aSig >= 0x01000000) { |
| 2568 |
aSig >>= 1;
|
| 2569 |
aExp++; |
| 2570 |
} |
| 2571 |
} else if (aExp < -14 |
| 2572 |
&& STATUS(float_detect_tininess) == float_tininess_before_rounding) {
|
| 2573 |
float_raise( float_flag_underflow STATUS_VAR); |
| 2574 |
} |
| 2575 |
|
| 2576 |
if (ieee) {
|
| 2577 |
if (aExp > 15) { |
| 2578 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 2579 |
return packFloat16(aSign, 0x1f, 0); |
| 2580 |
} |
| 2581 |
} else {
|
| 2582 |
if (aExp > 16) { |
| 2583 |
float_raise( float_flag_overflow | float_flag_inexact STATUS_VAR); |
| 2584 |
return packFloat16(aSign, 0x1f, 0x3ff); |
| 2585 |
} |
| 2586 |
} |
| 2587 |
if (aExp < -24) { |
| 2588 |
return packFloat16(aSign, 0, 0); |
| 2589 |
} |
| 2590 |
if (aExp < -14) { |
| 2591 |
aSig >>= -14 - aExp;
|
| 2592 |
aExp = -14;
|
| 2593 |
} |
| 2594 |
return packFloat16(aSign, aExp + 14, aSig >> 13); |
| 2595 |
} |
| 2596 |
|
| 2597 |
#ifdef FLOATX80
|
| 2598 |
|
| 2599 |
/*----------------------------------------------------------------------------
|
| 2600 |
| Returns the result of converting the double-precision floating-point value
|
| 2601 |
| `a' to the extended double-precision floating-point format. The conversion
|
| 2602 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2603 |
| Arithmetic.
|
| 2604 |
*----------------------------------------------------------------------------*/
|
| 2605 |
|
| 2606 |
floatx80 float64_to_floatx80( float64 a STATUS_PARAM ) |
| 2607 |
{
|
| 2608 |
flag aSign; |
| 2609 |
int16 aExp; |
| 2610 |
bits64 aSig; |
| 2611 |
|
| 2612 |
aSig = extractFloat64Frac( a ); |
| 2613 |
aExp = extractFloat64Exp( a ); |
| 2614 |
aSign = extractFloat64Sign( a ); |
| 2615 |
if ( aExp == 0x7FF ) { |
| 2616 |
if ( aSig ) return commonNaNToFloatx80( float64ToCommonNaN( a STATUS_VAR ) ); |
| 2617 |
return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 2618 |
} |
| 2619 |
if ( aExp == 0 ) { |
| 2620 |
if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); |
| 2621 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 2622 |
} |
| 2623 |
return
|
| 2624 |
packFloatx80( |
| 2625 |
aSign, aExp + 0x3C00, ( aSig | LIT64( 0x0010000000000000 ) )<<11 ); |
| 2626 |
|
| 2627 |
} |
| 2628 |
|
| 2629 |
#endif
|
| 2630 |
|
| 2631 |
#ifdef FLOAT128
|
| 2632 |
|
| 2633 |
/*----------------------------------------------------------------------------
|
| 2634 |
| Returns the result of converting the double-precision floating-point value
|
| 2635 |
| `a' to the quadruple-precision floating-point format. The conversion is
|
| 2636 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 2637 |
| Arithmetic.
|
| 2638 |
*----------------------------------------------------------------------------*/
|
| 2639 |
|
| 2640 |
float128 float64_to_float128( float64 a STATUS_PARAM ) |
| 2641 |
{
|
| 2642 |
flag aSign; |
| 2643 |
int16 aExp; |
| 2644 |
bits64 aSig, zSig0, zSig1; |
| 2645 |
|
| 2646 |
aSig = extractFloat64Frac( a ); |
| 2647 |
aExp = extractFloat64Exp( a ); |
| 2648 |
aSign = extractFloat64Sign( a ); |
| 2649 |
if ( aExp == 0x7FF ) { |
| 2650 |
if ( aSig ) return commonNaNToFloat128( float64ToCommonNaN( a STATUS_VAR ) ); |
| 2651 |
return packFloat128( aSign, 0x7FFF, 0, 0 ); |
| 2652 |
} |
| 2653 |
if ( aExp == 0 ) { |
| 2654 |
if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); |
| 2655 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 2656 |
--aExp; |
| 2657 |
} |
| 2658 |
shift128Right( aSig, 0, 4, &zSig0, &zSig1 ); |
| 2659 |
return packFloat128( aSign, aExp + 0x3C00, zSig0, zSig1 ); |
| 2660 |
|
| 2661 |
} |
| 2662 |
|
| 2663 |
#endif
|
| 2664 |
|
| 2665 |
/*----------------------------------------------------------------------------
|
| 2666 |
| Rounds the double-precision floating-point value `a' to an integer, and
|
| 2667 |
| returns the result as a double-precision floating-point value. The
|
| 2668 |
| operation is performed according to the IEC/IEEE Standard for Binary
|
| 2669 |
| Floating-Point Arithmetic.
|
| 2670 |
*----------------------------------------------------------------------------*/
|
| 2671 |
|
| 2672 |
float64 float64_round_to_int( float64 a STATUS_PARAM ) |
| 2673 |
{
|
| 2674 |
flag aSign; |
| 2675 |
int16 aExp; |
| 2676 |
bits64 lastBitMask, roundBitsMask; |
| 2677 |
int8 roundingMode; |
| 2678 |
bits64 z; |
| 2679 |
|
| 2680 |
aExp = extractFloat64Exp( a ); |
| 2681 |
if ( 0x433 <= aExp ) { |
| 2682 |
if ( ( aExp == 0x7FF ) && extractFloat64Frac( a ) ) { |
| 2683 |
return propagateFloat64NaN( a, a STATUS_VAR );
|
| 2684 |
} |
| 2685 |
return a;
|
| 2686 |
} |
| 2687 |
if ( aExp < 0x3FF ) { |
| 2688 |
if ( (bits64) ( float64_val(a)<<1 ) == 0 ) return a; |
| 2689 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 2690 |
aSign = extractFloat64Sign( a ); |
| 2691 |
switch ( STATUS(float_rounding_mode) ) {
|
| 2692 |
case float_round_nearest_even:
|
| 2693 |
if ( ( aExp == 0x3FE ) && extractFloat64Frac( a ) ) { |
| 2694 |
return packFloat64( aSign, 0x3FF, 0 ); |
| 2695 |
} |
| 2696 |
break;
|
| 2697 |
case float_round_down:
|
| 2698 |
return make_float64(aSign ? LIT64( 0xBFF0000000000000 ) : 0); |
| 2699 |
case float_round_up:
|
| 2700 |
return make_float64(
|
| 2701 |
aSign ? LIT64( 0x8000000000000000 ) : LIT64( 0x3FF0000000000000 )); |
| 2702 |
} |
| 2703 |
return packFloat64( aSign, 0, 0 ); |
| 2704 |
} |
| 2705 |
lastBitMask = 1;
|
| 2706 |
lastBitMask <<= 0x433 - aExp;
|
| 2707 |
roundBitsMask = lastBitMask - 1;
|
| 2708 |
z = float64_val(a); |
| 2709 |
roundingMode = STATUS(float_rounding_mode); |
| 2710 |
if ( roundingMode == float_round_nearest_even ) {
|
| 2711 |
z += lastBitMask>>1;
|
| 2712 |
if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; |
| 2713 |
} |
| 2714 |
else if ( roundingMode != float_round_to_zero ) { |
| 2715 |
if ( extractFloat64Sign( make_float64(z) ) ^ ( roundingMode == float_round_up ) ) {
|
| 2716 |
z += roundBitsMask; |
| 2717 |
} |
| 2718 |
} |
| 2719 |
z &= ~ roundBitsMask; |
| 2720 |
if ( z != float64_val(a) )
|
| 2721 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 2722 |
return make_float64(z);
|
| 2723 |
|
| 2724 |
} |
| 2725 |
|
| 2726 |
float64 float64_trunc_to_int( float64 a STATUS_PARAM) |
| 2727 |
{
|
| 2728 |
int oldmode;
|
| 2729 |
float64 res; |
| 2730 |
oldmode = STATUS(float_rounding_mode); |
| 2731 |
STATUS(float_rounding_mode) = float_round_to_zero; |
| 2732 |
res = float64_round_to_int(a STATUS_VAR); |
| 2733 |
STATUS(float_rounding_mode) = oldmode; |
| 2734 |
return res;
|
| 2735 |
} |
| 2736 |
|
| 2737 |
/*----------------------------------------------------------------------------
|
| 2738 |
| Returns the result of adding the absolute values of the double-precision
|
| 2739 |
| floating-point values `a' and `b'. If `zSign' is 1, the sum is negated
|
| 2740 |
| before being returned. `zSign' is ignored if the result is a NaN.
|
| 2741 |
| The addition is performed according to the IEC/IEEE Standard for Binary
|
| 2742 |
| Floating-Point Arithmetic.
|
| 2743 |
*----------------------------------------------------------------------------*/
|
| 2744 |
|
| 2745 |
static float64 addFloat64Sigs( float64 a, float64 b, flag zSign STATUS_PARAM )
|
| 2746 |
{
|
| 2747 |
int16 aExp, bExp, zExp; |
| 2748 |
bits64 aSig, bSig, zSig; |
| 2749 |
int16 expDiff; |
| 2750 |
|
| 2751 |
aSig = extractFloat64Frac( a ); |
| 2752 |
aExp = extractFloat64Exp( a ); |
| 2753 |
bSig = extractFloat64Frac( b ); |
| 2754 |
bExp = extractFloat64Exp( b ); |
| 2755 |
expDiff = aExp - bExp; |
| 2756 |
aSig <<= 9;
|
| 2757 |
bSig <<= 9;
|
| 2758 |
if ( 0 < expDiff ) { |
| 2759 |
if ( aExp == 0x7FF ) { |
| 2760 |
if ( aSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2761 |
return a;
|
| 2762 |
} |
| 2763 |
if ( bExp == 0 ) { |
| 2764 |
--expDiff; |
| 2765 |
} |
| 2766 |
else {
|
| 2767 |
bSig |= LIT64( 0x2000000000000000 );
|
| 2768 |
} |
| 2769 |
shift64RightJamming( bSig, expDiff, &bSig ); |
| 2770 |
zExp = aExp; |
| 2771 |
} |
| 2772 |
else if ( expDiff < 0 ) { |
| 2773 |
if ( bExp == 0x7FF ) { |
| 2774 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2775 |
return packFloat64( zSign, 0x7FF, 0 ); |
| 2776 |
} |
| 2777 |
if ( aExp == 0 ) { |
| 2778 |
++expDiff; |
| 2779 |
} |
| 2780 |
else {
|
| 2781 |
aSig |= LIT64( 0x2000000000000000 );
|
| 2782 |
} |
| 2783 |
shift64RightJamming( aSig, - expDiff, &aSig ); |
| 2784 |
zExp = bExp; |
| 2785 |
} |
| 2786 |
else {
|
| 2787 |
if ( aExp == 0x7FF ) { |
| 2788 |
if ( aSig | bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2789 |
return a;
|
| 2790 |
} |
| 2791 |
if ( aExp == 0 ) { |
| 2792 |
if ( STATUS(flush_to_zero) ) return packFloat64( zSign, 0, 0 ); |
| 2793 |
return packFloat64( zSign, 0, ( aSig + bSig )>>9 ); |
| 2794 |
} |
| 2795 |
zSig = LIT64( 0x4000000000000000 ) + aSig + bSig;
|
| 2796 |
zExp = aExp; |
| 2797 |
goto roundAndPack;
|
| 2798 |
} |
| 2799 |
aSig |= LIT64( 0x2000000000000000 );
|
| 2800 |
zSig = ( aSig + bSig )<<1;
|
| 2801 |
--zExp; |
| 2802 |
if ( (sbits64) zSig < 0 ) { |
| 2803 |
zSig = aSig + bSig; |
| 2804 |
++zExp; |
| 2805 |
} |
| 2806 |
roundAndPack:
|
| 2807 |
return roundAndPackFloat64( zSign, zExp, zSig STATUS_VAR );
|
| 2808 |
|
| 2809 |
} |
| 2810 |
|
| 2811 |
/*----------------------------------------------------------------------------
|
| 2812 |
| Returns the result of subtracting the absolute values of the double-
|
| 2813 |
| precision floating-point values `a' and `b'. If `zSign' is 1, the
|
| 2814 |
| difference is negated before being returned. `zSign' is ignored if the
|
| 2815 |
| result is a NaN. The subtraction is performed according to the IEC/IEEE
|
| 2816 |
| Standard for Binary Floating-Point Arithmetic.
|
| 2817 |
*----------------------------------------------------------------------------*/
|
| 2818 |
|
| 2819 |
static float64 subFloat64Sigs( float64 a, float64 b, flag zSign STATUS_PARAM )
|
| 2820 |
{
|
| 2821 |
int16 aExp, bExp, zExp; |
| 2822 |
bits64 aSig, bSig, zSig; |
| 2823 |
int16 expDiff; |
| 2824 |
|
| 2825 |
aSig = extractFloat64Frac( a ); |
| 2826 |
aExp = extractFloat64Exp( a ); |
| 2827 |
bSig = extractFloat64Frac( b ); |
| 2828 |
bExp = extractFloat64Exp( b ); |
| 2829 |
expDiff = aExp - bExp; |
| 2830 |
aSig <<= 10;
|
| 2831 |
bSig <<= 10;
|
| 2832 |
if ( 0 < expDiff ) goto aExpBigger; |
| 2833 |
if ( expDiff < 0 ) goto bExpBigger; |
| 2834 |
if ( aExp == 0x7FF ) { |
| 2835 |
if ( aSig | bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2836 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2837 |
return float64_default_nan;
|
| 2838 |
} |
| 2839 |
if ( aExp == 0 ) { |
| 2840 |
aExp = 1;
|
| 2841 |
bExp = 1;
|
| 2842 |
} |
| 2843 |
if ( bSig < aSig ) goto aBigger; |
| 2844 |
if ( aSig < bSig ) goto bBigger; |
| 2845 |
return packFloat64( STATUS(float_rounding_mode) == float_round_down, 0, 0 ); |
| 2846 |
bExpBigger:
|
| 2847 |
if ( bExp == 0x7FF ) { |
| 2848 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2849 |
return packFloat64( zSign ^ 1, 0x7FF, 0 ); |
| 2850 |
} |
| 2851 |
if ( aExp == 0 ) { |
| 2852 |
++expDiff; |
| 2853 |
} |
| 2854 |
else {
|
| 2855 |
aSig |= LIT64( 0x4000000000000000 );
|
| 2856 |
} |
| 2857 |
shift64RightJamming( aSig, - expDiff, &aSig ); |
| 2858 |
bSig |= LIT64( 0x4000000000000000 );
|
| 2859 |
bBigger:
|
| 2860 |
zSig = bSig - aSig; |
| 2861 |
zExp = bExp; |
| 2862 |
zSign ^= 1;
|
| 2863 |
goto normalizeRoundAndPack;
|
| 2864 |
aExpBigger:
|
| 2865 |
if ( aExp == 0x7FF ) { |
| 2866 |
if ( aSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2867 |
return a;
|
| 2868 |
} |
| 2869 |
if ( bExp == 0 ) { |
| 2870 |
--expDiff; |
| 2871 |
} |
| 2872 |
else {
|
| 2873 |
bSig |= LIT64( 0x4000000000000000 );
|
| 2874 |
} |
| 2875 |
shift64RightJamming( bSig, expDiff, &bSig ); |
| 2876 |
aSig |= LIT64( 0x4000000000000000 );
|
| 2877 |
aBigger:
|
| 2878 |
zSig = aSig - bSig; |
| 2879 |
zExp = aExp; |
| 2880 |
normalizeRoundAndPack:
|
| 2881 |
--zExp; |
| 2882 |
return normalizeRoundAndPackFloat64( zSign, zExp, zSig STATUS_VAR );
|
| 2883 |
|
| 2884 |
} |
| 2885 |
|
| 2886 |
/*----------------------------------------------------------------------------
|
| 2887 |
| Returns the result of adding the double-precision floating-point values `a'
|
| 2888 |
| and `b'. The operation is performed according to the IEC/IEEE Standard for
|
| 2889 |
| Binary Floating-Point Arithmetic.
|
| 2890 |
*----------------------------------------------------------------------------*/
|
| 2891 |
|
| 2892 |
float64 float64_add( float64 a, float64 b STATUS_PARAM ) |
| 2893 |
{
|
| 2894 |
flag aSign, bSign; |
| 2895 |
|
| 2896 |
aSign = extractFloat64Sign( a ); |
| 2897 |
bSign = extractFloat64Sign( b ); |
| 2898 |
if ( aSign == bSign ) {
|
| 2899 |
return addFloat64Sigs( a, b, aSign STATUS_VAR );
|
| 2900 |
} |
| 2901 |
else {
|
| 2902 |
return subFloat64Sigs( a, b, aSign STATUS_VAR );
|
| 2903 |
} |
| 2904 |
|
| 2905 |
} |
| 2906 |
|
| 2907 |
/*----------------------------------------------------------------------------
|
| 2908 |
| Returns the result of subtracting the double-precision floating-point values
|
| 2909 |
| `a' and `b'. The operation is performed according to the IEC/IEEE Standard
|
| 2910 |
| for Binary Floating-Point Arithmetic.
|
| 2911 |
*----------------------------------------------------------------------------*/
|
| 2912 |
|
| 2913 |
float64 float64_sub( float64 a, float64 b STATUS_PARAM ) |
| 2914 |
{
|
| 2915 |
flag aSign, bSign; |
| 2916 |
|
| 2917 |
aSign = extractFloat64Sign( a ); |
| 2918 |
bSign = extractFloat64Sign( b ); |
| 2919 |
if ( aSign == bSign ) {
|
| 2920 |
return subFloat64Sigs( a, b, aSign STATUS_VAR );
|
| 2921 |
} |
| 2922 |
else {
|
| 2923 |
return addFloat64Sigs( a, b, aSign STATUS_VAR );
|
| 2924 |
} |
| 2925 |
|
| 2926 |
} |
| 2927 |
|
| 2928 |
/*----------------------------------------------------------------------------
|
| 2929 |
| Returns the result of multiplying the double-precision floating-point values
|
| 2930 |
| `a' and `b'. The operation is performed according to the IEC/IEEE Standard
|
| 2931 |
| for Binary Floating-Point Arithmetic.
|
| 2932 |
*----------------------------------------------------------------------------*/
|
| 2933 |
|
| 2934 |
float64 float64_mul( float64 a, float64 b STATUS_PARAM ) |
| 2935 |
{
|
| 2936 |
flag aSign, bSign, zSign; |
| 2937 |
int16 aExp, bExp, zExp; |
| 2938 |
bits64 aSig, bSig, zSig0, zSig1; |
| 2939 |
|
| 2940 |
aSig = extractFloat64Frac( a ); |
| 2941 |
aExp = extractFloat64Exp( a ); |
| 2942 |
aSign = extractFloat64Sign( a ); |
| 2943 |
bSig = extractFloat64Frac( b ); |
| 2944 |
bExp = extractFloat64Exp( b ); |
| 2945 |
bSign = extractFloat64Sign( b ); |
| 2946 |
zSign = aSign ^ bSign; |
| 2947 |
if ( aExp == 0x7FF ) { |
| 2948 |
if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { |
| 2949 |
return propagateFloat64NaN( a, b STATUS_VAR );
|
| 2950 |
} |
| 2951 |
if ( ( bExp | bSig ) == 0 ) { |
| 2952 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2953 |
return float64_default_nan;
|
| 2954 |
} |
| 2955 |
return packFloat64( zSign, 0x7FF, 0 ); |
| 2956 |
} |
| 2957 |
if ( bExp == 0x7FF ) { |
| 2958 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 2959 |
if ( ( aExp | aSig ) == 0 ) { |
| 2960 |
float_raise( float_flag_invalid STATUS_VAR); |
| 2961 |
return float64_default_nan;
|
| 2962 |
} |
| 2963 |
return packFloat64( zSign, 0x7FF, 0 ); |
| 2964 |
} |
| 2965 |
if ( aExp == 0 ) { |
| 2966 |
if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); |
| 2967 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 2968 |
} |
| 2969 |
if ( bExp == 0 ) { |
| 2970 |
if ( bSig == 0 ) return packFloat64( zSign, 0, 0 ); |
| 2971 |
normalizeFloat64Subnormal( bSig, &bExp, &bSig ); |
| 2972 |
} |
| 2973 |
zExp = aExp + bExp - 0x3FF;
|
| 2974 |
aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; |
| 2975 |
bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; |
| 2976 |
mul64To128( aSig, bSig, &zSig0, &zSig1 ); |
| 2977 |
zSig0 |= ( zSig1 != 0 );
|
| 2978 |
if ( 0 <= (sbits64) ( zSig0<<1 ) ) { |
| 2979 |
zSig0 <<= 1;
|
| 2980 |
--zExp; |
| 2981 |
} |
| 2982 |
return roundAndPackFloat64( zSign, zExp, zSig0 STATUS_VAR );
|
| 2983 |
|
| 2984 |
} |
| 2985 |
|
| 2986 |
/*----------------------------------------------------------------------------
|
| 2987 |
| Returns the result of dividing the double-precision floating-point value `a'
|
| 2988 |
| by the corresponding value `b'. The operation is performed according to
|
| 2989 |
| the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 2990 |
*----------------------------------------------------------------------------*/
|
| 2991 |
|
| 2992 |
float64 float64_div( float64 a, float64 b STATUS_PARAM ) |
| 2993 |
{
|
| 2994 |
flag aSign, bSign, zSign; |
| 2995 |
int16 aExp, bExp, zExp; |
| 2996 |
bits64 aSig, bSig, zSig; |
| 2997 |
bits64 rem0, rem1; |
| 2998 |
bits64 term0, term1; |
| 2999 |
|
| 3000 |
aSig = extractFloat64Frac( a ); |
| 3001 |
aExp = extractFloat64Exp( a ); |
| 3002 |
aSign = extractFloat64Sign( a ); |
| 3003 |
bSig = extractFloat64Frac( b ); |
| 3004 |
bExp = extractFloat64Exp( b ); |
| 3005 |
bSign = extractFloat64Sign( b ); |
| 3006 |
zSign = aSign ^ bSign; |
| 3007 |
if ( aExp == 0x7FF ) { |
| 3008 |
if ( aSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 3009 |
if ( bExp == 0x7FF ) { |
| 3010 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 3011 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3012 |
return float64_default_nan;
|
| 3013 |
} |
| 3014 |
return packFloat64( zSign, 0x7FF, 0 ); |
| 3015 |
} |
| 3016 |
if ( bExp == 0x7FF ) { |
| 3017 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 3018 |
return packFloat64( zSign, 0, 0 ); |
| 3019 |
} |
| 3020 |
if ( bExp == 0 ) { |
| 3021 |
if ( bSig == 0 ) { |
| 3022 |
if ( ( aExp | aSig ) == 0 ) { |
| 3023 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3024 |
return float64_default_nan;
|
| 3025 |
} |
| 3026 |
float_raise( float_flag_divbyzero STATUS_VAR); |
| 3027 |
return packFloat64( zSign, 0x7FF, 0 ); |
| 3028 |
} |
| 3029 |
normalizeFloat64Subnormal( bSig, &bExp, &bSig ); |
| 3030 |
} |
| 3031 |
if ( aExp == 0 ) { |
| 3032 |
if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); |
| 3033 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 3034 |
} |
| 3035 |
zExp = aExp - bExp + 0x3FD;
|
| 3036 |
aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; |
| 3037 |
bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; |
| 3038 |
if ( bSig <= ( aSig + aSig ) ) {
|
| 3039 |
aSig >>= 1;
|
| 3040 |
++zExp; |
| 3041 |
} |
| 3042 |
zSig = estimateDiv128To64( aSig, 0, bSig );
|
| 3043 |
if ( ( zSig & 0x1FF ) <= 2 ) { |
| 3044 |
mul64To128( bSig, zSig, &term0, &term1 ); |
| 3045 |
sub128( aSig, 0, term0, term1, &rem0, &rem1 );
|
| 3046 |
while ( (sbits64) rem0 < 0 ) { |
| 3047 |
--zSig; |
| 3048 |
add128( rem0, rem1, 0, bSig, &rem0, &rem1 );
|
| 3049 |
} |
| 3050 |
zSig |= ( rem1 != 0 );
|
| 3051 |
} |
| 3052 |
return roundAndPackFloat64( zSign, zExp, zSig STATUS_VAR );
|
| 3053 |
|
| 3054 |
} |
| 3055 |
|
| 3056 |
/*----------------------------------------------------------------------------
|
| 3057 |
| Returns the remainder of the double-precision floating-point value `a'
|
| 3058 |
| with respect to the corresponding value `b'. The operation is performed
|
| 3059 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3060 |
*----------------------------------------------------------------------------*/
|
| 3061 |
|
| 3062 |
float64 float64_rem( float64 a, float64 b STATUS_PARAM ) |
| 3063 |
{
|
| 3064 |
flag aSign, zSign; |
| 3065 |
int16 aExp, bExp, expDiff; |
| 3066 |
bits64 aSig, bSig; |
| 3067 |
bits64 q, alternateASig; |
| 3068 |
sbits64 sigMean; |
| 3069 |
|
| 3070 |
aSig = extractFloat64Frac( a ); |
| 3071 |
aExp = extractFloat64Exp( a ); |
| 3072 |
aSign = extractFloat64Sign( a ); |
| 3073 |
bSig = extractFloat64Frac( b ); |
| 3074 |
bExp = extractFloat64Exp( b ); |
| 3075 |
if ( aExp == 0x7FF ) { |
| 3076 |
if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { |
| 3077 |
return propagateFloat64NaN( a, b STATUS_VAR );
|
| 3078 |
} |
| 3079 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3080 |
return float64_default_nan;
|
| 3081 |
} |
| 3082 |
if ( bExp == 0x7FF ) { |
| 3083 |
if ( bSig ) return propagateFloat64NaN( a, b STATUS_VAR ); |
| 3084 |
return a;
|
| 3085 |
} |
| 3086 |
if ( bExp == 0 ) { |
| 3087 |
if ( bSig == 0 ) { |
| 3088 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3089 |
return float64_default_nan;
|
| 3090 |
} |
| 3091 |
normalizeFloat64Subnormal( bSig, &bExp, &bSig ); |
| 3092 |
} |
| 3093 |
if ( aExp == 0 ) { |
| 3094 |
if ( aSig == 0 ) return a; |
| 3095 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 3096 |
} |
| 3097 |
expDiff = aExp - bExp; |
| 3098 |
aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<11; |
| 3099 |
bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; |
| 3100 |
if ( expDiff < 0 ) { |
| 3101 |
if ( expDiff < -1 ) return a; |
| 3102 |
aSig >>= 1;
|
| 3103 |
} |
| 3104 |
q = ( bSig <= aSig ); |
| 3105 |
if ( q ) aSig -= bSig;
|
| 3106 |
expDiff -= 64;
|
| 3107 |
while ( 0 < expDiff ) { |
| 3108 |
q = estimateDiv128To64( aSig, 0, bSig );
|
| 3109 |
q = ( 2 < q ) ? q - 2 : 0; |
| 3110 |
aSig = - ( ( bSig>>2 ) * q );
|
| 3111 |
expDiff -= 62;
|
| 3112 |
} |
| 3113 |
expDiff += 64;
|
| 3114 |
if ( 0 < expDiff ) { |
| 3115 |
q = estimateDiv128To64( aSig, 0, bSig );
|
| 3116 |
q = ( 2 < q ) ? q - 2 : 0; |
| 3117 |
q >>= 64 - expDiff;
|
| 3118 |
bSig >>= 2;
|
| 3119 |
aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; |
| 3120 |
} |
| 3121 |
else {
|
| 3122 |
aSig >>= 2;
|
| 3123 |
bSig >>= 2;
|
| 3124 |
} |
| 3125 |
do {
|
| 3126 |
alternateASig = aSig; |
| 3127 |
++q; |
| 3128 |
aSig -= bSig; |
| 3129 |
} while ( 0 <= (sbits64) aSig ); |
| 3130 |
sigMean = aSig + alternateASig; |
| 3131 |
if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { |
| 3132 |
aSig = alternateASig; |
| 3133 |
} |
| 3134 |
zSign = ( (sbits64) aSig < 0 );
|
| 3135 |
if ( zSign ) aSig = - aSig;
|
| 3136 |
return normalizeRoundAndPackFloat64( aSign ^ zSign, bExp, aSig STATUS_VAR );
|
| 3137 |
|
| 3138 |
} |
| 3139 |
|
| 3140 |
/*----------------------------------------------------------------------------
|
| 3141 |
| Returns the square root of the double-precision floating-point value `a'.
|
| 3142 |
| The operation is performed according to the IEC/IEEE Standard for Binary
|
| 3143 |
| Floating-Point Arithmetic.
|
| 3144 |
*----------------------------------------------------------------------------*/
|
| 3145 |
|
| 3146 |
float64 float64_sqrt( float64 a STATUS_PARAM ) |
| 3147 |
{
|
| 3148 |
flag aSign; |
| 3149 |
int16 aExp, zExp; |
| 3150 |
bits64 aSig, zSig, doubleZSig; |
| 3151 |
bits64 rem0, rem1, term0, term1; |
| 3152 |
|
| 3153 |
aSig = extractFloat64Frac( a ); |
| 3154 |
aExp = extractFloat64Exp( a ); |
| 3155 |
aSign = extractFloat64Sign( a ); |
| 3156 |
if ( aExp == 0x7FF ) { |
| 3157 |
if ( aSig ) return propagateFloat64NaN( a, a STATUS_VAR ); |
| 3158 |
if ( ! aSign ) return a; |
| 3159 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3160 |
return float64_default_nan;
|
| 3161 |
} |
| 3162 |
if ( aSign ) {
|
| 3163 |
if ( ( aExp | aSig ) == 0 ) return a; |
| 3164 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3165 |
return float64_default_nan;
|
| 3166 |
} |
| 3167 |
if ( aExp == 0 ) { |
| 3168 |
if ( aSig == 0 ) return float64_zero; |
| 3169 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 3170 |
} |
| 3171 |
zExp = ( ( aExp - 0x3FF )>>1 ) + 0x3FE; |
| 3172 |
aSig |= LIT64( 0x0010000000000000 );
|
| 3173 |
zSig = estimateSqrt32( aExp, aSig>>21 );
|
| 3174 |
aSig <<= 9 - ( aExp & 1 ); |
| 3175 |
zSig = estimateDiv128To64( aSig, 0, zSig<<32 ) + ( zSig<<30 ); |
| 3176 |
if ( ( zSig & 0x1FF ) <= 5 ) { |
| 3177 |
doubleZSig = zSig<<1;
|
| 3178 |
mul64To128( zSig, zSig, &term0, &term1 ); |
| 3179 |
sub128( aSig, 0, term0, term1, &rem0, &rem1 );
|
| 3180 |
while ( (sbits64) rem0 < 0 ) { |
| 3181 |
--zSig; |
| 3182 |
doubleZSig -= 2;
|
| 3183 |
add128( rem0, rem1, zSig>>63, doubleZSig | 1, &rem0, &rem1 ); |
| 3184 |
} |
| 3185 |
zSig |= ( ( rem0 | rem1 ) != 0 );
|
| 3186 |
} |
| 3187 |
return roundAndPackFloat64( 0, zExp, zSig STATUS_VAR ); |
| 3188 |
|
| 3189 |
} |
| 3190 |
|
| 3191 |
/*----------------------------------------------------------------------------
|
| 3192 |
| Returns the binary log of the double-precision floating-point value `a'.
|
| 3193 |
| The operation is performed according to the IEC/IEEE Standard for Binary
|
| 3194 |
| Floating-Point Arithmetic.
|
| 3195 |
*----------------------------------------------------------------------------*/
|
| 3196 |
float64 float64_log2( float64 a STATUS_PARAM ) |
| 3197 |
{
|
| 3198 |
flag aSign, zSign; |
| 3199 |
int16 aExp; |
| 3200 |
bits64 aSig, aSig0, aSig1, zSig, i; |
| 3201 |
|
| 3202 |
aSig = extractFloat64Frac( a ); |
| 3203 |
aExp = extractFloat64Exp( a ); |
| 3204 |
aSign = extractFloat64Sign( a ); |
| 3205 |
|
| 3206 |
if ( aExp == 0 ) { |
| 3207 |
if ( aSig == 0 ) return packFloat64( 1, 0x7FF, 0 ); |
| 3208 |
normalizeFloat64Subnormal( aSig, &aExp, &aSig ); |
| 3209 |
} |
| 3210 |
if ( aSign ) {
|
| 3211 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3212 |
return float64_default_nan;
|
| 3213 |
} |
| 3214 |
if ( aExp == 0x7FF ) { |
| 3215 |
if ( aSig ) return propagateFloat64NaN( a, float64_zero STATUS_VAR ); |
| 3216 |
return a;
|
| 3217 |
} |
| 3218 |
|
| 3219 |
aExp -= 0x3FF;
|
| 3220 |
aSig |= LIT64( 0x0010000000000000 );
|
| 3221 |
zSign = aExp < 0;
|
| 3222 |
zSig = (bits64)aExp << 52;
|
| 3223 |
for (i = 1LL << 51; i > 0; i >>= 1) { |
| 3224 |
mul64To128( aSig, aSig, &aSig0, &aSig1 ); |
| 3225 |
aSig = ( aSig0 << 12 ) | ( aSig1 >> 52 ); |
| 3226 |
if ( aSig & LIT64( 0x0020000000000000 ) ) { |
| 3227 |
aSig >>= 1;
|
| 3228 |
zSig |= i; |
| 3229 |
} |
| 3230 |
} |
| 3231 |
|
| 3232 |
if ( zSign )
|
| 3233 |
zSig = -zSig; |
| 3234 |
return normalizeRoundAndPackFloat64( zSign, 0x408, zSig STATUS_VAR ); |
| 3235 |
} |
| 3236 |
|
| 3237 |
/*----------------------------------------------------------------------------
|
| 3238 |
| Returns 1 if the double-precision floating-point value `a' is equal to the
|
| 3239 |
| corresponding value `b', and 0 otherwise. The comparison is performed
|
| 3240 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3241 |
*----------------------------------------------------------------------------*/
|
| 3242 |
|
| 3243 |
int float64_eq( float64 a, float64 b STATUS_PARAM )
|
| 3244 |
{
|
| 3245 |
bits64 av, bv; |
| 3246 |
|
| 3247 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3248 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3249 |
) {
|
| 3250 |
if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) {
|
| 3251 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3252 |
} |
| 3253 |
return 0; |
| 3254 |
} |
| 3255 |
av = float64_val(a); |
| 3256 |
bv = float64_val(b); |
| 3257 |
return ( av == bv ) || ( (bits64) ( ( av | bv )<<1 ) == 0 ); |
| 3258 |
|
| 3259 |
} |
| 3260 |
|
| 3261 |
/*----------------------------------------------------------------------------
|
| 3262 |
| Returns 1 if the double-precision floating-point value `a' is less than or
|
| 3263 |
| equal to the corresponding value `b', and 0 otherwise. The comparison is
|
| 3264 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 3265 |
| Arithmetic.
|
| 3266 |
*----------------------------------------------------------------------------*/
|
| 3267 |
|
| 3268 |
int float64_le( float64 a, float64 b STATUS_PARAM )
|
| 3269 |
{
|
| 3270 |
flag aSign, bSign; |
| 3271 |
bits64 av, bv; |
| 3272 |
|
| 3273 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3274 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3275 |
) {
|
| 3276 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3277 |
return 0; |
| 3278 |
} |
| 3279 |
aSign = extractFloat64Sign( a ); |
| 3280 |
bSign = extractFloat64Sign( b ); |
| 3281 |
av = float64_val(a); |
| 3282 |
bv = float64_val(b); |
| 3283 |
if ( aSign != bSign ) return aSign || ( (bits64) ( ( av | bv )<<1 ) == 0 ); |
| 3284 |
return ( av == bv ) || ( aSign ^ ( av < bv ) );
|
| 3285 |
|
| 3286 |
} |
| 3287 |
|
| 3288 |
/*----------------------------------------------------------------------------
|
| 3289 |
| Returns 1 if the double-precision floating-point value `a' is less than
|
| 3290 |
| the corresponding value `b', and 0 otherwise. The comparison is performed
|
| 3291 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3292 |
*----------------------------------------------------------------------------*/
|
| 3293 |
|
| 3294 |
int float64_lt( float64 a, float64 b STATUS_PARAM )
|
| 3295 |
{
|
| 3296 |
flag aSign, bSign; |
| 3297 |
bits64 av, bv; |
| 3298 |
|
| 3299 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3300 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3301 |
) {
|
| 3302 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3303 |
return 0; |
| 3304 |
} |
| 3305 |
aSign = extractFloat64Sign( a ); |
| 3306 |
bSign = extractFloat64Sign( b ); |
| 3307 |
av = float64_val(a); |
| 3308 |
bv = float64_val(b); |
| 3309 |
if ( aSign != bSign ) return aSign && ( (bits64) ( ( av | bv )<<1 ) != 0 ); |
| 3310 |
return ( av != bv ) && ( aSign ^ ( av < bv ) );
|
| 3311 |
|
| 3312 |
} |
| 3313 |
|
| 3314 |
/*----------------------------------------------------------------------------
|
| 3315 |
| Returns 1 if the double-precision floating-point value `a' is equal to the
|
| 3316 |
| corresponding value `b', and 0 otherwise. The invalid exception is raised
|
| 3317 |
| if either operand is a NaN. Otherwise, the comparison is performed
|
| 3318 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3319 |
*----------------------------------------------------------------------------*/
|
| 3320 |
|
| 3321 |
int float64_eq_signaling( float64 a, float64 b STATUS_PARAM )
|
| 3322 |
{
|
| 3323 |
bits64 av, bv; |
| 3324 |
|
| 3325 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3326 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3327 |
) {
|
| 3328 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3329 |
return 0; |
| 3330 |
} |
| 3331 |
av = float64_val(a); |
| 3332 |
bv = float64_val(b); |
| 3333 |
return ( av == bv ) || ( (bits64) ( ( av | bv )<<1 ) == 0 ); |
| 3334 |
|
| 3335 |
} |
| 3336 |
|
| 3337 |
/*----------------------------------------------------------------------------
|
| 3338 |
| Returns 1 if the double-precision floating-point value `a' is less than or
|
| 3339 |
| equal to the corresponding value `b', and 0 otherwise. Quiet NaNs do not
|
| 3340 |
| cause an exception. Otherwise, the comparison is performed according to the
|
| 3341 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3342 |
*----------------------------------------------------------------------------*/
|
| 3343 |
|
| 3344 |
int float64_le_quiet( float64 a, float64 b STATUS_PARAM )
|
| 3345 |
{
|
| 3346 |
flag aSign, bSign; |
| 3347 |
bits64 av, bv; |
| 3348 |
|
| 3349 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3350 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3351 |
) {
|
| 3352 |
if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) {
|
| 3353 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3354 |
} |
| 3355 |
return 0; |
| 3356 |
} |
| 3357 |
aSign = extractFloat64Sign( a ); |
| 3358 |
bSign = extractFloat64Sign( b ); |
| 3359 |
av = float64_val(a); |
| 3360 |
bv = float64_val(b); |
| 3361 |
if ( aSign != bSign ) return aSign || ( (bits64) ( ( av | bv )<<1 ) == 0 ); |
| 3362 |
return ( av == bv ) || ( aSign ^ ( av < bv ) );
|
| 3363 |
|
| 3364 |
} |
| 3365 |
|
| 3366 |
/*----------------------------------------------------------------------------
|
| 3367 |
| Returns 1 if the double-precision floating-point value `a' is less than
|
| 3368 |
| the corresponding value `b', and 0 otherwise. Quiet NaNs do not cause an
|
| 3369 |
| exception. Otherwise, the comparison is performed according to the IEC/IEEE
|
| 3370 |
| Standard for Binary Floating-Point Arithmetic.
|
| 3371 |
*----------------------------------------------------------------------------*/
|
| 3372 |
|
| 3373 |
int float64_lt_quiet( float64 a, float64 b STATUS_PARAM )
|
| 3374 |
{
|
| 3375 |
flag aSign, bSign; |
| 3376 |
bits64 av, bv; |
| 3377 |
|
| 3378 |
if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) |
| 3379 |
|| ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) )
|
| 3380 |
) {
|
| 3381 |
if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) {
|
| 3382 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3383 |
} |
| 3384 |
return 0; |
| 3385 |
} |
| 3386 |
aSign = extractFloat64Sign( a ); |
| 3387 |
bSign = extractFloat64Sign( b ); |
| 3388 |
av = float64_val(a); |
| 3389 |
bv = float64_val(b); |
| 3390 |
if ( aSign != bSign ) return aSign && ( (bits64) ( ( av | bv )<<1 ) != 0 ); |
| 3391 |
return ( av != bv ) && ( aSign ^ ( av < bv ) );
|
| 3392 |
|
| 3393 |
} |
| 3394 |
|
| 3395 |
#ifdef FLOATX80
|
| 3396 |
|
| 3397 |
/*----------------------------------------------------------------------------
|
| 3398 |
| Returns the result of converting the extended double-precision floating-
|
| 3399 |
| point value `a' to the 32-bit two's complement integer format. The
|
| 3400 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3401 |
| Floating-Point Arithmetic---which means in particular that the conversion
|
| 3402 |
| is rounded according to the current rounding mode. If `a' is a NaN, the
|
| 3403 |
| largest positive integer is returned. Otherwise, if the conversion
|
| 3404 |
| overflows, the largest integer with the same sign as `a' is returned.
|
| 3405 |
*----------------------------------------------------------------------------*/
|
| 3406 |
|
| 3407 |
int32 floatx80_to_int32( floatx80 a STATUS_PARAM ) |
| 3408 |
{
|
| 3409 |
flag aSign; |
| 3410 |
int32 aExp, shiftCount; |
| 3411 |
bits64 aSig; |
| 3412 |
|
| 3413 |
aSig = extractFloatx80Frac( a ); |
| 3414 |
aExp = extractFloatx80Exp( a ); |
| 3415 |
aSign = extractFloatx80Sign( a ); |
| 3416 |
if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; |
| 3417 |
shiftCount = 0x4037 - aExp;
|
| 3418 |
if ( shiftCount <= 0 ) shiftCount = 1; |
| 3419 |
shift64RightJamming( aSig, shiftCount, &aSig ); |
| 3420 |
return roundAndPackInt32( aSign, aSig STATUS_VAR );
|
| 3421 |
|
| 3422 |
} |
| 3423 |
|
| 3424 |
/*----------------------------------------------------------------------------
|
| 3425 |
| Returns the result of converting the extended double-precision floating-
|
| 3426 |
| point value `a' to the 32-bit two's complement integer format. The
|
| 3427 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3428 |
| Floating-Point Arithmetic, except that the conversion is always rounded
|
| 3429 |
| toward zero. If `a' is a NaN, the largest positive integer is returned.
|
| 3430 |
| Otherwise, if the conversion overflows, the largest integer with the same
|
| 3431 |
| sign as `a' is returned.
|
| 3432 |
*----------------------------------------------------------------------------*/
|
| 3433 |
|
| 3434 |
int32 floatx80_to_int32_round_to_zero( floatx80 a STATUS_PARAM ) |
| 3435 |
{
|
| 3436 |
flag aSign; |
| 3437 |
int32 aExp, shiftCount; |
| 3438 |
bits64 aSig, savedASig; |
| 3439 |
int32 z; |
| 3440 |
|
| 3441 |
aSig = extractFloatx80Frac( a ); |
| 3442 |
aExp = extractFloatx80Exp( a ); |
| 3443 |
aSign = extractFloatx80Sign( a ); |
| 3444 |
if ( 0x401E < aExp ) { |
| 3445 |
if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; |
| 3446 |
goto invalid;
|
| 3447 |
} |
| 3448 |
else if ( aExp < 0x3FFF ) { |
| 3449 |
if ( aExp || aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 3450 |
return 0; |
| 3451 |
} |
| 3452 |
shiftCount = 0x403E - aExp;
|
| 3453 |
savedASig = aSig; |
| 3454 |
aSig >>= shiftCount; |
| 3455 |
z = aSig; |
| 3456 |
if ( aSign ) z = - z;
|
| 3457 |
if ( ( z < 0 ) ^ aSign ) { |
| 3458 |
invalid:
|
| 3459 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3460 |
return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; |
| 3461 |
} |
| 3462 |
if ( ( aSig<<shiftCount ) != savedASig ) {
|
| 3463 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 3464 |
} |
| 3465 |
return z;
|
| 3466 |
|
| 3467 |
} |
| 3468 |
|
| 3469 |
/*----------------------------------------------------------------------------
|
| 3470 |
| Returns the result of converting the extended double-precision floating-
|
| 3471 |
| point value `a' to the 64-bit two's complement integer format. The
|
| 3472 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3473 |
| Floating-Point Arithmetic---which means in particular that the conversion
|
| 3474 |
| is rounded according to the current rounding mode. If `a' is a NaN,
|
| 3475 |
| the largest positive integer is returned. Otherwise, if the conversion
|
| 3476 |
| overflows, the largest integer with the same sign as `a' is returned.
|
| 3477 |
*----------------------------------------------------------------------------*/
|
| 3478 |
|
| 3479 |
int64 floatx80_to_int64( floatx80 a STATUS_PARAM ) |
| 3480 |
{
|
| 3481 |
flag aSign; |
| 3482 |
int32 aExp, shiftCount; |
| 3483 |
bits64 aSig, aSigExtra; |
| 3484 |
|
| 3485 |
aSig = extractFloatx80Frac( a ); |
| 3486 |
aExp = extractFloatx80Exp( a ); |
| 3487 |
aSign = extractFloatx80Sign( a ); |
| 3488 |
shiftCount = 0x403E - aExp;
|
| 3489 |
if ( shiftCount <= 0 ) { |
| 3490 |
if ( shiftCount ) {
|
| 3491 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3492 |
if ( ! aSign
|
| 3493 |
|| ( ( aExp == 0x7FFF )
|
| 3494 |
&& ( aSig != LIT64( 0x8000000000000000 ) ) )
|
| 3495 |
) {
|
| 3496 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 3497 |
} |
| 3498 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 3499 |
} |
| 3500 |
aSigExtra = 0;
|
| 3501 |
} |
| 3502 |
else {
|
| 3503 |
shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra );
|
| 3504 |
} |
| 3505 |
return roundAndPackInt64( aSign, aSig, aSigExtra STATUS_VAR );
|
| 3506 |
|
| 3507 |
} |
| 3508 |
|
| 3509 |
/*----------------------------------------------------------------------------
|
| 3510 |
| Returns the result of converting the extended double-precision floating-
|
| 3511 |
| point value `a' to the 64-bit two's complement integer format. The
|
| 3512 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3513 |
| Floating-Point Arithmetic, except that the conversion is always rounded
|
| 3514 |
| toward zero. If `a' is a NaN, the largest positive integer is returned.
|
| 3515 |
| Otherwise, if the conversion overflows, the largest integer with the same
|
| 3516 |
| sign as `a' is returned.
|
| 3517 |
*----------------------------------------------------------------------------*/
|
| 3518 |
|
| 3519 |
int64 floatx80_to_int64_round_to_zero( floatx80 a STATUS_PARAM ) |
| 3520 |
{
|
| 3521 |
flag aSign; |
| 3522 |
int32 aExp, shiftCount; |
| 3523 |
bits64 aSig; |
| 3524 |
int64 z; |
| 3525 |
|
| 3526 |
aSig = extractFloatx80Frac( a ); |
| 3527 |
aExp = extractFloatx80Exp( a ); |
| 3528 |
aSign = extractFloatx80Sign( a ); |
| 3529 |
shiftCount = aExp - 0x403E;
|
| 3530 |
if ( 0 <= shiftCount ) { |
| 3531 |
aSig &= LIT64( 0x7FFFFFFFFFFFFFFF );
|
| 3532 |
if ( ( a.high != 0xC03E ) || aSig ) { |
| 3533 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3534 |
if ( ! aSign || ( ( aExp == 0x7FFF ) && aSig ) ) { |
| 3535 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 3536 |
} |
| 3537 |
} |
| 3538 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 3539 |
} |
| 3540 |
else if ( aExp < 0x3FFF ) { |
| 3541 |
if ( aExp | aSig ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 3542 |
return 0; |
| 3543 |
} |
| 3544 |
z = aSig>>( - shiftCount ); |
| 3545 |
if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { |
| 3546 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 3547 |
} |
| 3548 |
if ( aSign ) z = - z;
|
| 3549 |
return z;
|
| 3550 |
|
| 3551 |
} |
| 3552 |
|
| 3553 |
/*----------------------------------------------------------------------------
|
| 3554 |
| Returns the result of converting the extended double-precision floating-
|
| 3555 |
| point value `a' to the single-precision floating-point format. The
|
| 3556 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3557 |
| Floating-Point Arithmetic.
|
| 3558 |
*----------------------------------------------------------------------------*/
|
| 3559 |
|
| 3560 |
float32 floatx80_to_float32( floatx80 a STATUS_PARAM ) |
| 3561 |
{
|
| 3562 |
flag aSign; |
| 3563 |
int32 aExp; |
| 3564 |
bits64 aSig; |
| 3565 |
|
| 3566 |
aSig = extractFloatx80Frac( a ); |
| 3567 |
aExp = extractFloatx80Exp( a ); |
| 3568 |
aSign = extractFloatx80Sign( a ); |
| 3569 |
if ( aExp == 0x7FFF ) { |
| 3570 |
if ( (bits64) ( aSig<<1 ) ) { |
| 3571 |
return commonNaNToFloat32( floatx80ToCommonNaN( a STATUS_VAR ) );
|
| 3572 |
} |
| 3573 |
return packFloat32( aSign, 0xFF, 0 ); |
| 3574 |
} |
| 3575 |
shift64RightJamming( aSig, 33, &aSig );
|
| 3576 |
if ( aExp || aSig ) aExp -= 0x3F81; |
| 3577 |
return roundAndPackFloat32( aSign, aExp, aSig STATUS_VAR );
|
| 3578 |
|
| 3579 |
} |
| 3580 |
|
| 3581 |
/*----------------------------------------------------------------------------
|
| 3582 |
| Returns the result of converting the extended double-precision floating-
|
| 3583 |
| point value `a' to the double-precision floating-point format. The
|
| 3584 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3585 |
| Floating-Point Arithmetic.
|
| 3586 |
*----------------------------------------------------------------------------*/
|
| 3587 |
|
| 3588 |
float64 floatx80_to_float64( floatx80 a STATUS_PARAM ) |
| 3589 |
{
|
| 3590 |
flag aSign; |
| 3591 |
int32 aExp; |
| 3592 |
bits64 aSig, zSig; |
| 3593 |
|
| 3594 |
aSig = extractFloatx80Frac( a ); |
| 3595 |
aExp = extractFloatx80Exp( a ); |
| 3596 |
aSign = extractFloatx80Sign( a ); |
| 3597 |
if ( aExp == 0x7FFF ) { |
| 3598 |
if ( (bits64) ( aSig<<1 ) ) { |
| 3599 |
return commonNaNToFloat64( floatx80ToCommonNaN( a STATUS_VAR ) );
|
| 3600 |
} |
| 3601 |
return packFloat64( aSign, 0x7FF, 0 ); |
| 3602 |
} |
| 3603 |
shift64RightJamming( aSig, 1, &zSig );
|
| 3604 |
if ( aExp || aSig ) aExp -= 0x3C01; |
| 3605 |
return roundAndPackFloat64( aSign, aExp, zSig STATUS_VAR );
|
| 3606 |
|
| 3607 |
} |
| 3608 |
|
| 3609 |
#ifdef FLOAT128
|
| 3610 |
|
| 3611 |
/*----------------------------------------------------------------------------
|
| 3612 |
| Returns the result of converting the extended double-precision floating-
|
| 3613 |
| point value `a' to the quadruple-precision floating-point format. The
|
| 3614 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 3615 |
| Floating-Point Arithmetic.
|
| 3616 |
*----------------------------------------------------------------------------*/
|
| 3617 |
|
| 3618 |
float128 floatx80_to_float128( floatx80 a STATUS_PARAM ) |
| 3619 |
{
|
| 3620 |
flag aSign; |
| 3621 |
int16 aExp; |
| 3622 |
bits64 aSig, zSig0, zSig1; |
| 3623 |
|
| 3624 |
aSig = extractFloatx80Frac( a ); |
| 3625 |
aExp = extractFloatx80Exp( a ); |
| 3626 |
aSign = extractFloatx80Sign( a ); |
| 3627 |
if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) { |
| 3628 |
return commonNaNToFloat128( floatx80ToCommonNaN( a STATUS_VAR ) );
|
| 3629 |
} |
| 3630 |
shift128Right( aSig<<1, 0, 16, &zSig0, &zSig1 ); |
| 3631 |
return packFloat128( aSign, aExp, zSig0, zSig1 );
|
| 3632 |
|
| 3633 |
} |
| 3634 |
|
| 3635 |
#endif
|
| 3636 |
|
| 3637 |
/*----------------------------------------------------------------------------
|
| 3638 |
| Rounds the extended double-precision floating-point value `a' to an integer,
|
| 3639 |
| and returns the result as an extended quadruple-precision floating-point
|
| 3640 |
| value. The operation is performed according to the IEC/IEEE Standard for
|
| 3641 |
| Binary Floating-Point Arithmetic.
|
| 3642 |
*----------------------------------------------------------------------------*/
|
| 3643 |
|
| 3644 |
floatx80 floatx80_round_to_int( floatx80 a STATUS_PARAM ) |
| 3645 |
{
|
| 3646 |
flag aSign; |
| 3647 |
int32 aExp; |
| 3648 |
bits64 lastBitMask, roundBitsMask; |
| 3649 |
int8 roundingMode; |
| 3650 |
floatx80 z; |
| 3651 |
|
| 3652 |
aExp = extractFloatx80Exp( a ); |
| 3653 |
if ( 0x403E <= aExp ) { |
| 3654 |
if ( ( aExp == 0x7FFF ) && (bits64) ( extractFloatx80Frac( a )<<1 ) ) { |
| 3655 |
return propagateFloatx80NaN( a, a STATUS_VAR );
|
| 3656 |
} |
| 3657 |
return a;
|
| 3658 |
} |
| 3659 |
if ( aExp < 0x3FFF ) { |
| 3660 |
if ( ( aExp == 0 ) |
| 3661 |
&& ( (bits64) ( extractFloatx80Frac( a )<<1 ) == 0 ) ) { |
| 3662 |
return a;
|
| 3663 |
} |
| 3664 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 3665 |
aSign = extractFloatx80Sign( a ); |
| 3666 |
switch ( STATUS(float_rounding_mode) ) {
|
| 3667 |
case float_round_nearest_even:
|
| 3668 |
if ( ( aExp == 0x3FFE ) && (bits64) ( extractFloatx80Frac( a )<<1 ) |
| 3669 |
) {
|
| 3670 |
return
|
| 3671 |
packFloatx80( aSign, 0x3FFF, LIT64( 0x8000000000000000 ) ); |
| 3672 |
} |
| 3673 |
break;
|
| 3674 |
case float_round_down:
|
| 3675 |
return
|
| 3676 |
aSign ? |
| 3677 |
packFloatx80( 1, 0x3FFF, LIT64( 0x8000000000000000 ) ) |
| 3678 |
: packFloatx80( 0, 0, 0 ); |
| 3679 |
case float_round_up:
|
| 3680 |
return
|
| 3681 |
aSign ? packFloatx80( 1, 0, 0 ) |
| 3682 |
: packFloatx80( 0, 0x3FFF, LIT64( 0x8000000000000000 ) ); |
| 3683 |
} |
| 3684 |
return packFloatx80( aSign, 0, 0 ); |
| 3685 |
} |
| 3686 |
lastBitMask = 1;
|
| 3687 |
lastBitMask <<= 0x403E - aExp;
|
| 3688 |
roundBitsMask = lastBitMask - 1;
|
| 3689 |
z = a; |
| 3690 |
roundingMode = STATUS(float_rounding_mode); |
| 3691 |
if ( roundingMode == float_round_nearest_even ) {
|
| 3692 |
z.low += lastBitMask>>1;
|
| 3693 |
if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; |
| 3694 |
} |
| 3695 |
else if ( roundingMode != float_round_to_zero ) { |
| 3696 |
if ( extractFloatx80Sign( z ) ^ ( roundingMode == float_round_up ) ) {
|
| 3697 |
z.low += roundBitsMask; |
| 3698 |
} |
| 3699 |
} |
| 3700 |
z.low &= ~ roundBitsMask; |
| 3701 |
if ( z.low == 0 ) { |
| 3702 |
++z.high; |
| 3703 |
z.low = LIT64( 0x8000000000000000 );
|
| 3704 |
} |
| 3705 |
if ( z.low != a.low ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 3706 |
return z;
|
| 3707 |
|
| 3708 |
} |
| 3709 |
|
| 3710 |
/*----------------------------------------------------------------------------
|
| 3711 |
| Returns the result of adding the absolute values of the extended double-
|
| 3712 |
| precision floating-point values `a' and `b'. If `zSign' is 1, the sum is
|
| 3713 |
| negated before being returned. `zSign' is ignored if the result is a NaN.
|
| 3714 |
| The addition is performed according to the IEC/IEEE Standard for Binary
|
| 3715 |
| Floating-Point Arithmetic.
|
| 3716 |
*----------------------------------------------------------------------------*/
|
| 3717 |
|
| 3718 |
static floatx80 addFloatx80Sigs( floatx80 a, floatx80 b, flag zSign STATUS_PARAM)
|
| 3719 |
{
|
| 3720 |
int32 aExp, bExp, zExp; |
| 3721 |
bits64 aSig, bSig, zSig0, zSig1; |
| 3722 |
int32 expDiff; |
| 3723 |
|
| 3724 |
aSig = extractFloatx80Frac( a ); |
| 3725 |
aExp = extractFloatx80Exp( a ); |
| 3726 |
bSig = extractFloatx80Frac( b ); |
| 3727 |
bExp = extractFloatx80Exp( b ); |
| 3728 |
expDiff = aExp - bExp; |
| 3729 |
if ( 0 < expDiff ) { |
| 3730 |
if ( aExp == 0x7FFF ) { |
| 3731 |
if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3732 |
return a;
|
| 3733 |
} |
| 3734 |
if ( bExp == 0 ) --expDiff; |
| 3735 |
shift64ExtraRightJamming( bSig, 0, expDiff, &bSig, &zSig1 );
|
| 3736 |
zExp = aExp; |
| 3737 |
} |
| 3738 |
else if ( expDiff < 0 ) { |
| 3739 |
if ( bExp == 0x7FFF ) { |
| 3740 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3741 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3742 |
} |
| 3743 |
if ( aExp == 0 ) ++expDiff; |
| 3744 |
shift64ExtraRightJamming( aSig, 0, - expDiff, &aSig, &zSig1 );
|
| 3745 |
zExp = bExp; |
| 3746 |
} |
| 3747 |
else {
|
| 3748 |
if ( aExp == 0x7FFF ) { |
| 3749 |
if ( (bits64) ( ( aSig | bSig )<<1 ) ) { |
| 3750 |
return propagateFloatx80NaN( a, b STATUS_VAR );
|
| 3751 |
} |
| 3752 |
return a;
|
| 3753 |
} |
| 3754 |
zSig1 = 0;
|
| 3755 |
zSig0 = aSig + bSig; |
| 3756 |
if ( aExp == 0 ) { |
| 3757 |
normalizeFloatx80Subnormal( zSig0, &zExp, &zSig0 ); |
| 3758 |
goto roundAndPack;
|
| 3759 |
} |
| 3760 |
zExp = aExp; |
| 3761 |
goto shiftRight1;
|
| 3762 |
} |
| 3763 |
zSig0 = aSig + bSig; |
| 3764 |
if ( (sbits64) zSig0 < 0 ) goto roundAndPack; |
| 3765 |
shiftRight1:
|
| 3766 |
shift64ExtraRightJamming( zSig0, zSig1, 1, &zSig0, &zSig1 );
|
| 3767 |
zSig0 |= LIT64( 0x8000000000000000 );
|
| 3768 |
++zExp; |
| 3769 |
roundAndPack:
|
| 3770 |
return
|
| 3771 |
roundAndPackFloatx80( |
| 3772 |
STATUS(floatx80_rounding_precision), zSign, zExp, zSig0, zSig1 STATUS_VAR ); |
| 3773 |
|
| 3774 |
} |
| 3775 |
|
| 3776 |
/*----------------------------------------------------------------------------
|
| 3777 |
| Returns the result of subtracting the absolute values of the extended
|
| 3778 |
| double-precision floating-point values `a' and `b'. If `zSign' is 1, the
|
| 3779 |
| difference is negated before being returned. `zSign' is ignored if the
|
| 3780 |
| result is a NaN. The subtraction is performed according to the IEC/IEEE
|
| 3781 |
| Standard for Binary Floating-Point Arithmetic.
|
| 3782 |
*----------------------------------------------------------------------------*/
|
| 3783 |
|
| 3784 |
static floatx80 subFloatx80Sigs( floatx80 a, floatx80 b, flag zSign STATUS_PARAM )
|
| 3785 |
{
|
| 3786 |
int32 aExp, bExp, zExp; |
| 3787 |
bits64 aSig, bSig, zSig0, zSig1; |
| 3788 |
int32 expDiff; |
| 3789 |
floatx80 z; |
| 3790 |
|
| 3791 |
aSig = extractFloatx80Frac( a ); |
| 3792 |
aExp = extractFloatx80Exp( a ); |
| 3793 |
bSig = extractFloatx80Frac( b ); |
| 3794 |
bExp = extractFloatx80Exp( b ); |
| 3795 |
expDiff = aExp - bExp; |
| 3796 |
if ( 0 < expDiff ) goto aExpBigger; |
| 3797 |
if ( expDiff < 0 ) goto bExpBigger; |
| 3798 |
if ( aExp == 0x7FFF ) { |
| 3799 |
if ( (bits64) ( ( aSig | bSig )<<1 ) ) { |
| 3800 |
return propagateFloatx80NaN( a, b STATUS_VAR );
|
| 3801 |
} |
| 3802 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3803 |
z.low = floatx80_default_nan_low; |
| 3804 |
z.high = floatx80_default_nan_high; |
| 3805 |
return z;
|
| 3806 |
} |
| 3807 |
if ( aExp == 0 ) { |
| 3808 |
aExp = 1;
|
| 3809 |
bExp = 1;
|
| 3810 |
} |
| 3811 |
zSig1 = 0;
|
| 3812 |
if ( bSig < aSig ) goto aBigger; |
| 3813 |
if ( aSig < bSig ) goto bBigger; |
| 3814 |
return packFloatx80( STATUS(float_rounding_mode) == float_round_down, 0, 0 ); |
| 3815 |
bExpBigger:
|
| 3816 |
if ( bExp == 0x7FFF ) { |
| 3817 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3818 |
return packFloatx80( zSign ^ 1, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3819 |
} |
| 3820 |
if ( aExp == 0 ) ++expDiff; |
| 3821 |
shift128RightJamming( aSig, 0, - expDiff, &aSig, &zSig1 );
|
| 3822 |
bBigger:
|
| 3823 |
sub128( bSig, 0, aSig, zSig1, &zSig0, &zSig1 );
|
| 3824 |
zExp = bExp; |
| 3825 |
zSign ^= 1;
|
| 3826 |
goto normalizeRoundAndPack;
|
| 3827 |
aExpBigger:
|
| 3828 |
if ( aExp == 0x7FFF ) { |
| 3829 |
if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3830 |
return a;
|
| 3831 |
} |
| 3832 |
if ( bExp == 0 ) --expDiff; |
| 3833 |
shift128RightJamming( bSig, 0, expDiff, &bSig, &zSig1 );
|
| 3834 |
aBigger:
|
| 3835 |
sub128( aSig, 0, bSig, zSig1, &zSig0, &zSig1 );
|
| 3836 |
zExp = aExp; |
| 3837 |
normalizeRoundAndPack:
|
| 3838 |
return
|
| 3839 |
normalizeRoundAndPackFloatx80( |
| 3840 |
STATUS(floatx80_rounding_precision), zSign, zExp, zSig0, zSig1 STATUS_VAR ); |
| 3841 |
|
| 3842 |
} |
| 3843 |
|
| 3844 |
/*----------------------------------------------------------------------------
|
| 3845 |
| Returns the result of adding the extended double-precision floating-point
|
| 3846 |
| values `a' and `b'. The operation is performed according to the IEC/IEEE
|
| 3847 |
| Standard for Binary Floating-Point Arithmetic.
|
| 3848 |
*----------------------------------------------------------------------------*/
|
| 3849 |
|
| 3850 |
floatx80 floatx80_add( floatx80 a, floatx80 b STATUS_PARAM ) |
| 3851 |
{
|
| 3852 |
flag aSign, bSign; |
| 3853 |
|
| 3854 |
aSign = extractFloatx80Sign( a ); |
| 3855 |
bSign = extractFloatx80Sign( b ); |
| 3856 |
if ( aSign == bSign ) {
|
| 3857 |
return addFloatx80Sigs( a, b, aSign STATUS_VAR );
|
| 3858 |
} |
| 3859 |
else {
|
| 3860 |
return subFloatx80Sigs( a, b, aSign STATUS_VAR );
|
| 3861 |
} |
| 3862 |
|
| 3863 |
} |
| 3864 |
|
| 3865 |
/*----------------------------------------------------------------------------
|
| 3866 |
| Returns the result of subtracting the extended double-precision floating-
|
| 3867 |
| point values `a' and `b'. The operation is performed according to the
|
| 3868 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3869 |
*----------------------------------------------------------------------------*/
|
| 3870 |
|
| 3871 |
floatx80 floatx80_sub( floatx80 a, floatx80 b STATUS_PARAM ) |
| 3872 |
{
|
| 3873 |
flag aSign, bSign; |
| 3874 |
|
| 3875 |
aSign = extractFloatx80Sign( a ); |
| 3876 |
bSign = extractFloatx80Sign( b ); |
| 3877 |
if ( aSign == bSign ) {
|
| 3878 |
return subFloatx80Sigs( a, b, aSign STATUS_VAR );
|
| 3879 |
} |
| 3880 |
else {
|
| 3881 |
return addFloatx80Sigs( a, b, aSign STATUS_VAR );
|
| 3882 |
} |
| 3883 |
|
| 3884 |
} |
| 3885 |
|
| 3886 |
/*----------------------------------------------------------------------------
|
| 3887 |
| Returns the result of multiplying the extended double-precision floating-
|
| 3888 |
| point values `a' and `b'. The operation is performed according to the
|
| 3889 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3890 |
*----------------------------------------------------------------------------*/
|
| 3891 |
|
| 3892 |
floatx80 floatx80_mul( floatx80 a, floatx80 b STATUS_PARAM ) |
| 3893 |
{
|
| 3894 |
flag aSign, bSign, zSign; |
| 3895 |
int32 aExp, bExp, zExp; |
| 3896 |
bits64 aSig, bSig, zSig0, zSig1; |
| 3897 |
floatx80 z; |
| 3898 |
|
| 3899 |
aSig = extractFloatx80Frac( a ); |
| 3900 |
aExp = extractFloatx80Exp( a ); |
| 3901 |
aSign = extractFloatx80Sign( a ); |
| 3902 |
bSig = extractFloatx80Frac( b ); |
| 3903 |
bExp = extractFloatx80Exp( b ); |
| 3904 |
bSign = extractFloatx80Sign( b ); |
| 3905 |
zSign = aSign ^ bSign; |
| 3906 |
if ( aExp == 0x7FFF ) { |
| 3907 |
if ( (bits64) ( aSig<<1 ) |
| 3908 |
|| ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { |
| 3909 |
return propagateFloatx80NaN( a, b STATUS_VAR );
|
| 3910 |
} |
| 3911 |
if ( ( bExp | bSig ) == 0 ) goto invalid; |
| 3912 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3913 |
} |
| 3914 |
if ( bExp == 0x7FFF ) { |
| 3915 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3916 |
if ( ( aExp | aSig ) == 0 ) { |
| 3917 |
invalid:
|
| 3918 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3919 |
z.low = floatx80_default_nan_low; |
| 3920 |
z.high = floatx80_default_nan_high; |
| 3921 |
return z;
|
| 3922 |
} |
| 3923 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3924 |
} |
| 3925 |
if ( aExp == 0 ) { |
| 3926 |
if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); |
| 3927 |
normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); |
| 3928 |
} |
| 3929 |
if ( bExp == 0 ) { |
| 3930 |
if ( bSig == 0 ) return packFloatx80( zSign, 0, 0 ); |
| 3931 |
normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); |
| 3932 |
} |
| 3933 |
zExp = aExp + bExp - 0x3FFE;
|
| 3934 |
mul64To128( aSig, bSig, &zSig0, &zSig1 ); |
| 3935 |
if ( 0 < (sbits64) zSig0 ) { |
| 3936 |
shortShift128Left( zSig0, zSig1, 1, &zSig0, &zSig1 );
|
| 3937 |
--zExp; |
| 3938 |
} |
| 3939 |
return
|
| 3940 |
roundAndPackFloatx80( |
| 3941 |
STATUS(floatx80_rounding_precision), zSign, zExp, zSig0, zSig1 STATUS_VAR ); |
| 3942 |
|
| 3943 |
} |
| 3944 |
|
| 3945 |
/*----------------------------------------------------------------------------
|
| 3946 |
| Returns the result of dividing the extended double-precision floating-point
|
| 3947 |
| value `a' by the corresponding value `b'. The operation is performed
|
| 3948 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 3949 |
*----------------------------------------------------------------------------*/
|
| 3950 |
|
| 3951 |
floatx80 floatx80_div( floatx80 a, floatx80 b STATUS_PARAM ) |
| 3952 |
{
|
| 3953 |
flag aSign, bSign, zSign; |
| 3954 |
int32 aExp, bExp, zExp; |
| 3955 |
bits64 aSig, bSig, zSig0, zSig1; |
| 3956 |
bits64 rem0, rem1, rem2, term0, term1, term2; |
| 3957 |
floatx80 z; |
| 3958 |
|
| 3959 |
aSig = extractFloatx80Frac( a ); |
| 3960 |
aExp = extractFloatx80Exp( a ); |
| 3961 |
aSign = extractFloatx80Sign( a ); |
| 3962 |
bSig = extractFloatx80Frac( b ); |
| 3963 |
bExp = extractFloatx80Exp( b ); |
| 3964 |
bSign = extractFloatx80Sign( b ); |
| 3965 |
zSign = aSign ^ bSign; |
| 3966 |
if ( aExp == 0x7FFF ) { |
| 3967 |
if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3968 |
if ( bExp == 0x7FFF ) { |
| 3969 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3970 |
goto invalid;
|
| 3971 |
} |
| 3972 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3973 |
} |
| 3974 |
if ( bExp == 0x7FFF ) { |
| 3975 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 3976 |
return packFloatx80( zSign, 0, 0 ); |
| 3977 |
} |
| 3978 |
if ( bExp == 0 ) { |
| 3979 |
if ( bSig == 0 ) { |
| 3980 |
if ( ( aExp | aSig ) == 0 ) { |
| 3981 |
invalid:
|
| 3982 |
float_raise( float_flag_invalid STATUS_VAR); |
| 3983 |
z.low = floatx80_default_nan_low; |
| 3984 |
z.high = floatx80_default_nan_high; |
| 3985 |
return z;
|
| 3986 |
} |
| 3987 |
float_raise( float_flag_divbyzero STATUS_VAR); |
| 3988 |
return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 3989 |
} |
| 3990 |
normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); |
| 3991 |
} |
| 3992 |
if ( aExp == 0 ) { |
| 3993 |
if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); |
| 3994 |
normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); |
| 3995 |
} |
| 3996 |
zExp = aExp - bExp + 0x3FFE;
|
| 3997 |
rem1 = 0;
|
| 3998 |
if ( bSig <= aSig ) {
|
| 3999 |
shift128Right( aSig, 0, 1, &aSig, &rem1 ); |
| 4000 |
++zExp; |
| 4001 |
} |
| 4002 |
zSig0 = estimateDiv128To64( aSig, rem1, bSig ); |
| 4003 |
mul64To128( bSig, zSig0, &term0, &term1 ); |
| 4004 |
sub128( aSig, rem1, term0, term1, &rem0, &rem1 ); |
| 4005 |
while ( (sbits64) rem0 < 0 ) { |
| 4006 |
--zSig0; |
| 4007 |
add128( rem0, rem1, 0, bSig, &rem0, &rem1 );
|
| 4008 |
} |
| 4009 |
zSig1 = estimateDiv128To64( rem1, 0, bSig );
|
| 4010 |
if ( (bits64) ( zSig1<<1 ) <= 8 ) { |
| 4011 |
mul64To128( bSig, zSig1, &term1, &term2 ); |
| 4012 |
sub128( rem1, 0, term1, term2, &rem1, &rem2 );
|
| 4013 |
while ( (sbits64) rem1 < 0 ) { |
| 4014 |
--zSig1; |
| 4015 |
add128( rem1, rem2, 0, bSig, &rem1, &rem2 );
|
| 4016 |
} |
| 4017 |
zSig1 |= ( ( rem1 | rem2 ) != 0 );
|
| 4018 |
} |
| 4019 |
return
|
| 4020 |
roundAndPackFloatx80( |
| 4021 |
STATUS(floatx80_rounding_precision), zSign, zExp, zSig0, zSig1 STATUS_VAR ); |
| 4022 |
|
| 4023 |
} |
| 4024 |
|
| 4025 |
/*----------------------------------------------------------------------------
|
| 4026 |
| Returns the remainder of the extended double-precision floating-point value
|
| 4027 |
| `a' with respect to the corresponding value `b'. The operation is performed
|
| 4028 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 4029 |
*----------------------------------------------------------------------------*/
|
| 4030 |
|
| 4031 |
floatx80 floatx80_rem( floatx80 a, floatx80 b STATUS_PARAM ) |
| 4032 |
{
|
| 4033 |
flag aSign, zSign; |
| 4034 |
int32 aExp, bExp, expDiff; |
| 4035 |
bits64 aSig0, aSig1, bSig; |
| 4036 |
bits64 q, term0, term1, alternateASig0, alternateASig1; |
| 4037 |
floatx80 z; |
| 4038 |
|
| 4039 |
aSig0 = extractFloatx80Frac( a ); |
| 4040 |
aExp = extractFloatx80Exp( a ); |
| 4041 |
aSign = extractFloatx80Sign( a ); |
| 4042 |
bSig = extractFloatx80Frac( b ); |
| 4043 |
bExp = extractFloatx80Exp( b ); |
| 4044 |
if ( aExp == 0x7FFF ) { |
| 4045 |
if ( (bits64) ( aSig0<<1 ) |
| 4046 |
|| ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { |
| 4047 |
return propagateFloatx80NaN( a, b STATUS_VAR );
|
| 4048 |
} |
| 4049 |
goto invalid;
|
| 4050 |
} |
| 4051 |
if ( bExp == 0x7FFF ) { |
| 4052 |
if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b STATUS_VAR ); |
| 4053 |
return a;
|
| 4054 |
} |
| 4055 |
if ( bExp == 0 ) { |
| 4056 |
if ( bSig == 0 ) { |
| 4057 |
invalid:
|
| 4058 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4059 |
z.low = floatx80_default_nan_low; |
| 4060 |
z.high = floatx80_default_nan_high; |
| 4061 |
return z;
|
| 4062 |
} |
| 4063 |
normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); |
| 4064 |
} |
| 4065 |
if ( aExp == 0 ) { |
| 4066 |
if ( (bits64) ( aSig0<<1 ) == 0 ) return a; |
| 4067 |
normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); |
| 4068 |
} |
| 4069 |
bSig |= LIT64( 0x8000000000000000 );
|
| 4070 |
zSign = aSign; |
| 4071 |
expDiff = aExp - bExp; |
| 4072 |
aSig1 = 0;
|
| 4073 |
if ( expDiff < 0 ) { |
| 4074 |
if ( expDiff < -1 ) return a; |
| 4075 |
shift128Right( aSig0, 0, 1, &aSig0, &aSig1 ); |
| 4076 |
expDiff = 0;
|
| 4077 |
} |
| 4078 |
q = ( bSig <= aSig0 ); |
| 4079 |
if ( q ) aSig0 -= bSig;
|
| 4080 |
expDiff -= 64;
|
| 4081 |
while ( 0 < expDiff ) { |
| 4082 |
q = estimateDiv128To64( aSig0, aSig1, bSig ); |
| 4083 |
q = ( 2 < q ) ? q - 2 : 0; |
| 4084 |
mul64To128( bSig, q, &term0, &term1 ); |
| 4085 |
sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); |
| 4086 |
shortShift128Left( aSig0, aSig1, 62, &aSig0, &aSig1 );
|
| 4087 |
expDiff -= 62;
|
| 4088 |
} |
| 4089 |
expDiff += 64;
|
| 4090 |
if ( 0 < expDiff ) { |
| 4091 |
q = estimateDiv128To64( aSig0, aSig1, bSig ); |
| 4092 |
q = ( 2 < q ) ? q - 2 : 0; |
| 4093 |
q >>= 64 - expDiff;
|
| 4094 |
mul64To128( bSig, q<<( 64 - expDiff ), &term0, &term1 );
|
| 4095 |
sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); |
| 4096 |
shortShift128Left( 0, bSig, 64 - expDiff, &term0, &term1 ); |
| 4097 |
while ( le128( term0, term1, aSig0, aSig1 ) ) {
|
| 4098 |
++q; |
| 4099 |
sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); |
| 4100 |
} |
| 4101 |
} |
| 4102 |
else {
|
| 4103 |
term1 = 0;
|
| 4104 |
term0 = bSig; |
| 4105 |
} |
| 4106 |
sub128( term0, term1, aSig0, aSig1, &alternateASig0, &alternateASig1 ); |
| 4107 |
if ( lt128( alternateASig0, alternateASig1, aSig0, aSig1 )
|
| 4108 |
|| ( eq128( alternateASig0, alternateASig1, aSig0, aSig1 ) |
| 4109 |
&& ( q & 1 ) )
|
| 4110 |
) {
|
| 4111 |
aSig0 = alternateASig0; |
| 4112 |
aSig1 = alternateASig1; |
| 4113 |
zSign = ! zSign; |
| 4114 |
} |
| 4115 |
return
|
| 4116 |
normalizeRoundAndPackFloatx80( |
| 4117 |
80, zSign, bExp + expDiff, aSig0, aSig1 STATUS_VAR );
|
| 4118 |
|
| 4119 |
} |
| 4120 |
|
| 4121 |
/*----------------------------------------------------------------------------
|
| 4122 |
| Returns the square root of the extended double-precision floating-point
|
| 4123 |
| value `a'. The operation is performed according to the IEC/IEEE Standard
|
| 4124 |
| for Binary Floating-Point Arithmetic.
|
| 4125 |
*----------------------------------------------------------------------------*/
|
| 4126 |
|
| 4127 |
floatx80 floatx80_sqrt( floatx80 a STATUS_PARAM ) |
| 4128 |
{
|
| 4129 |
flag aSign; |
| 4130 |
int32 aExp, zExp; |
| 4131 |
bits64 aSig0, aSig1, zSig0, zSig1, doubleZSig0; |
| 4132 |
bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; |
| 4133 |
floatx80 z; |
| 4134 |
|
| 4135 |
aSig0 = extractFloatx80Frac( a ); |
| 4136 |
aExp = extractFloatx80Exp( a ); |
| 4137 |
aSign = extractFloatx80Sign( a ); |
| 4138 |
if ( aExp == 0x7FFF ) { |
| 4139 |
if ( (bits64) ( aSig0<<1 ) ) return propagateFloatx80NaN( a, a STATUS_VAR ); |
| 4140 |
if ( ! aSign ) return a; |
| 4141 |
goto invalid;
|
| 4142 |
} |
| 4143 |
if ( aSign ) {
|
| 4144 |
if ( ( aExp | aSig0 ) == 0 ) return a; |
| 4145 |
invalid:
|
| 4146 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4147 |
z.low = floatx80_default_nan_low; |
| 4148 |
z.high = floatx80_default_nan_high; |
| 4149 |
return z;
|
| 4150 |
} |
| 4151 |
if ( aExp == 0 ) { |
| 4152 |
if ( aSig0 == 0 ) return packFloatx80( 0, 0, 0 ); |
| 4153 |
normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); |
| 4154 |
} |
| 4155 |
zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFF; |
| 4156 |
zSig0 = estimateSqrt32( aExp, aSig0>>32 );
|
| 4157 |
shift128Right( aSig0, 0, 2 + ( aExp & 1 ), &aSig0, &aSig1 ); |
| 4158 |
zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); |
| 4159 |
doubleZSig0 = zSig0<<1;
|
| 4160 |
mul64To128( zSig0, zSig0, &term0, &term1 ); |
| 4161 |
sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); |
| 4162 |
while ( (sbits64) rem0 < 0 ) { |
| 4163 |
--zSig0; |
| 4164 |
doubleZSig0 -= 2;
|
| 4165 |
add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); |
| 4166 |
} |
| 4167 |
zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 );
|
| 4168 |
if ( ( zSig1 & LIT64( 0x3FFFFFFFFFFFFFFF ) ) <= 5 ) { |
| 4169 |
if ( zSig1 == 0 ) zSig1 = 1; |
| 4170 |
mul64To128( doubleZSig0, zSig1, &term1, &term2 ); |
| 4171 |
sub128( rem1, 0, term1, term2, &rem1, &rem2 );
|
| 4172 |
mul64To128( zSig1, zSig1, &term2, &term3 ); |
| 4173 |
sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); |
| 4174 |
while ( (sbits64) rem1 < 0 ) { |
| 4175 |
--zSig1; |
| 4176 |
shortShift128Left( 0, zSig1, 1, &term2, &term3 ); |
| 4177 |
term3 |= 1;
|
| 4178 |
term2 |= doubleZSig0; |
| 4179 |
add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 );
|
| 4180 |
} |
| 4181 |
zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 );
|
| 4182 |
} |
| 4183 |
shortShift128Left( 0, zSig1, 1, &zSig0, &zSig1 ); |
| 4184 |
zSig0 |= doubleZSig0; |
| 4185 |
return
|
| 4186 |
roundAndPackFloatx80( |
| 4187 |
STATUS(floatx80_rounding_precision), 0, zExp, zSig0, zSig1 STATUS_VAR );
|
| 4188 |
|
| 4189 |
} |
| 4190 |
|
| 4191 |
/*----------------------------------------------------------------------------
|
| 4192 |
| Returns 1 if the extended double-precision floating-point value `a' is
|
| 4193 |
| equal to the corresponding value `b', and 0 otherwise. The comparison is
|
| 4194 |
| performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4195 |
| Arithmetic.
|
| 4196 |
*----------------------------------------------------------------------------*/
|
| 4197 |
|
| 4198 |
int floatx80_eq( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4199 |
{
|
| 4200 |
|
| 4201 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4202 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4203 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4204 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4205 |
) {
|
| 4206 |
if ( floatx80_is_signaling_nan( a )
|
| 4207 |
|| floatx80_is_signaling_nan( b ) ) {
|
| 4208 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4209 |
} |
| 4210 |
return 0; |
| 4211 |
} |
| 4212 |
return
|
| 4213 |
( a.low == b.low ) |
| 4214 |
&& ( ( a.high == b.high ) |
| 4215 |
|| ( ( a.low == 0 )
|
| 4216 |
&& ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) |
| 4217 |
); |
| 4218 |
|
| 4219 |
} |
| 4220 |
|
| 4221 |
/*----------------------------------------------------------------------------
|
| 4222 |
| Returns 1 if the extended double-precision floating-point value `a' is
|
| 4223 |
| less than or equal to the corresponding value `b', and 0 otherwise. The
|
| 4224 |
| comparison is performed according to the IEC/IEEE Standard for Binary
|
| 4225 |
| Floating-Point Arithmetic.
|
| 4226 |
*----------------------------------------------------------------------------*/
|
| 4227 |
|
| 4228 |
int floatx80_le( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4229 |
{
|
| 4230 |
flag aSign, bSign; |
| 4231 |
|
| 4232 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4233 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4234 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4235 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4236 |
) {
|
| 4237 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4238 |
return 0; |
| 4239 |
} |
| 4240 |
aSign = extractFloatx80Sign( a ); |
| 4241 |
bSign = extractFloatx80Sign( b ); |
| 4242 |
if ( aSign != bSign ) {
|
| 4243 |
return
|
| 4244 |
aSign |
| 4245 |
|| ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 4246 |
== 0 );
|
| 4247 |
} |
| 4248 |
return
|
| 4249 |
aSign ? le128( b.high, b.low, a.high, a.low ) |
| 4250 |
: le128( a.high, a.low, b.high, b.low ); |
| 4251 |
|
| 4252 |
} |
| 4253 |
|
| 4254 |
/*----------------------------------------------------------------------------
|
| 4255 |
| Returns 1 if the extended double-precision floating-point value `a' is
|
| 4256 |
| less than the corresponding value `b', and 0 otherwise. The comparison
|
| 4257 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4258 |
| Arithmetic.
|
| 4259 |
*----------------------------------------------------------------------------*/
|
| 4260 |
|
| 4261 |
int floatx80_lt( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4262 |
{
|
| 4263 |
flag aSign, bSign; |
| 4264 |
|
| 4265 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4266 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4267 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4268 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4269 |
) {
|
| 4270 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4271 |
return 0; |
| 4272 |
} |
| 4273 |
aSign = extractFloatx80Sign( a ); |
| 4274 |
bSign = extractFloatx80Sign( b ); |
| 4275 |
if ( aSign != bSign ) {
|
| 4276 |
return
|
| 4277 |
aSign |
| 4278 |
&& ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 4279 |
!= 0 );
|
| 4280 |
} |
| 4281 |
return
|
| 4282 |
aSign ? lt128( b.high, b.low, a.high, a.low ) |
| 4283 |
: lt128( a.high, a.low, b.high, b.low ); |
| 4284 |
|
| 4285 |
} |
| 4286 |
|
| 4287 |
/*----------------------------------------------------------------------------
|
| 4288 |
| Returns 1 if the extended double-precision floating-point value `a' is equal
|
| 4289 |
| to the corresponding value `b', and 0 otherwise. The invalid exception is
|
| 4290 |
| raised if either operand is a NaN. Otherwise, the comparison is performed
|
| 4291 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 4292 |
*----------------------------------------------------------------------------*/
|
| 4293 |
|
| 4294 |
int floatx80_eq_signaling( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4295 |
{
|
| 4296 |
|
| 4297 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4298 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4299 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4300 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4301 |
) {
|
| 4302 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4303 |
return 0; |
| 4304 |
} |
| 4305 |
return
|
| 4306 |
( a.low == b.low ) |
| 4307 |
&& ( ( a.high == b.high ) |
| 4308 |
|| ( ( a.low == 0 )
|
| 4309 |
&& ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) |
| 4310 |
); |
| 4311 |
|
| 4312 |
} |
| 4313 |
|
| 4314 |
/*----------------------------------------------------------------------------
|
| 4315 |
| Returns 1 if the extended double-precision floating-point value `a' is less
|
| 4316 |
| than or equal to the corresponding value `b', and 0 otherwise. Quiet NaNs
|
| 4317 |
| do not cause an exception. Otherwise, the comparison is performed according
|
| 4318 |
| to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 4319 |
*----------------------------------------------------------------------------*/
|
| 4320 |
|
| 4321 |
int floatx80_le_quiet( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4322 |
{
|
| 4323 |
flag aSign, bSign; |
| 4324 |
|
| 4325 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4326 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4327 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4328 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4329 |
) {
|
| 4330 |
if ( floatx80_is_signaling_nan( a )
|
| 4331 |
|| floatx80_is_signaling_nan( b ) ) {
|
| 4332 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4333 |
} |
| 4334 |
return 0; |
| 4335 |
} |
| 4336 |
aSign = extractFloatx80Sign( a ); |
| 4337 |
bSign = extractFloatx80Sign( b ); |
| 4338 |
if ( aSign != bSign ) {
|
| 4339 |
return
|
| 4340 |
aSign |
| 4341 |
|| ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 4342 |
== 0 );
|
| 4343 |
} |
| 4344 |
return
|
| 4345 |
aSign ? le128( b.high, b.low, a.high, a.low ) |
| 4346 |
: le128( a.high, a.low, b.high, b.low ); |
| 4347 |
|
| 4348 |
} |
| 4349 |
|
| 4350 |
/*----------------------------------------------------------------------------
|
| 4351 |
| Returns 1 if the extended double-precision floating-point value `a' is less
|
| 4352 |
| than the corresponding value `b', and 0 otherwise. Quiet NaNs do not cause
|
| 4353 |
| an exception. Otherwise, the comparison is performed according to the
|
| 4354 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 4355 |
*----------------------------------------------------------------------------*/
|
| 4356 |
|
| 4357 |
int floatx80_lt_quiet( floatx80 a, floatx80 b STATUS_PARAM )
|
| 4358 |
{
|
| 4359 |
flag aSign, bSign; |
| 4360 |
|
| 4361 |
if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) |
| 4362 |
&& (bits64) ( extractFloatx80Frac( a )<<1 ) )
|
| 4363 |
|| ( ( extractFloatx80Exp( b ) == 0x7FFF )
|
| 4364 |
&& (bits64) ( extractFloatx80Frac( b )<<1 ) )
|
| 4365 |
) {
|
| 4366 |
if ( floatx80_is_signaling_nan( a )
|
| 4367 |
|| floatx80_is_signaling_nan( b ) ) {
|
| 4368 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4369 |
} |
| 4370 |
return 0; |
| 4371 |
} |
| 4372 |
aSign = extractFloatx80Sign( a ); |
| 4373 |
bSign = extractFloatx80Sign( b ); |
| 4374 |
if ( aSign != bSign ) {
|
| 4375 |
return
|
| 4376 |
aSign |
| 4377 |
&& ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 4378 |
!= 0 );
|
| 4379 |
} |
| 4380 |
return
|
| 4381 |
aSign ? lt128( b.high, b.low, a.high, a.low ) |
| 4382 |
: lt128( a.high, a.low, b.high, b.low ); |
| 4383 |
|
| 4384 |
} |
| 4385 |
|
| 4386 |
#endif
|
| 4387 |
|
| 4388 |
#ifdef FLOAT128
|
| 4389 |
|
| 4390 |
/*----------------------------------------------------------------------------
|
| 4391 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4392 |
| value `a' to the 32-bit two's complement integer format. The conversion
|
| 4393 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4394 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 4395 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 4396 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 4397 |
| largest integer with the same sign as `a' is returned.
|
| 4398 |
*----------------------------------------------------------------------------*/
|
| 4399 |
|
| 4400 |
int32 float128_to_int32( float128 a STATUS_PARAM ) |
| 4401 |
{
|
| 4402 |
flag aSign; |
| 4403 |
int32 aExp, shiftCount; |
| 4404 |
bits64 aSig0, aSig1; |
| 4405 |
|
| 4406 |
aSig1 = extractFloat128Frac1( a ); |
| 4407 |
aSig0 = extractFloat128Frac0( a ); |
| 4408 |
aExp = extractFloat128Exp( a ); |
| 4409 |
aSign = extractFloat128Sign( a ); |
| 4410 |
if ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) aSign = 0; |
| 4411 |
if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); |
| 4412 |
aSig0 |= ( aSig1 != 0 );
|
| 4413 |
shiftCount = 0x4028 - aExp;
|
| 4414 |
if ( 0 < shiftCount ) shift64RightJamming( aSig0, shiftCount, &aSig0 ); |
| 4415 |
return roundAndPackInt32( aSign, aSig0 STATUS_VAR );
|
| 4416 |
|
| 4417 |
} |
| 4418 |
|
| 4419 |
/*----------------------------------------------------------------------------
|
| 4420 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4421 |
| value `a' to the 32-bit two's complement integer format. The conversion
|
| 4422 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4423 |
| Arithmetic, except that the conversion is always rounded toward zero. If
|
| 4424 |
| `a' is a NaN, the largest positive integer is returned. Otherwise, if the
|
| 4425 |
| conversion overflows, the largest integer with the same sign as `a' is
|
| 4426 |
| returned.
|
| 4427 |
*----------------------------------------------------------------------------*/
|
| 4428 |
|
| 4429 |
int32 float128_to_int32_round_to_zero( float128 a STATUS_PARAM ) |
| 4430 |
{
|
| 4431 |
flag aSign; |
| 4432 |
int32 aExp, shiftCount; |
| 4433 |
bits64 aSig0, aSig1, savedASig; |
| 4434 |
int32 z; |
| 4435 |
|
| 4436 |
aSig1 = extractFloat128Frac1( a ); |
| 4437 |
aSig0 = extractFloat128Frac0( a ); |
| 4438 |
aExp = extractFloat128Exp( a ); |
| 4439 |
aSign = extractFloat128Sign( a ); |
| 4440 |
aSig0 |= ( aSig1 != 0 );
|
| 4441 |
if ( 0x401E < aExp ) { |
| 4442 |
if ( ( aExp == 0x7FFF ) && aSig0 ) aSign = 0; |
| 4443 |
goto invalid;
|
| 4444 |
} |
| 4445 |
else if ( aExp < 0x3FFF ) { |
| 4446 |
if ( aExp || aSig0 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 4447 |
return 0; |
| 4448 |
} |
| 4449 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 4450 |
shiftCount = 0x402F - aExp;
|
| 4451 |
savedASig = aSig0; |
| 4452 |
aSig0 >>= shiftCount; |
| 4453 |
z = aSig0; |
| 4454 |
if ( aSign ) z = - z;
|
| 4455 |
if ( ( z < 0 ) ^ aSign ) { |
| 4456 |
invalid:
|
| 4457 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4458 |
return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; |
| 4459 |
} |
| 4460 |
if ( ( aSig0<<shiftCount ) != savedASig ) {
|
| 4461 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4462 |
} |
| 4463 |
return z;
|
| 4464 |
|
| 4465 |
} |
| 4466 |
|
| 4467 |
/*----------------------------------------------------------------------------
|
| 4468 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4469 |
| value `a' to the 64-bit two's complement integer format. The conversion
|
| 4470 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4471 |
| Arithmetic---which means in particular that the conversion is rounded
|
| 4472 |
| according to the current rounding mode. If `a' is a NaN, the largest
|
| 4473 |
| positive integer is returned. Otherwise, if the conversion overflows, the
|
| 4474 |
| largest integer with the same sign as `a' is returned.
|
| 4475 |
*----------------------------------------------------------------------------*/
|
| 4476 |
|
| 4477 |
int64 float128_to_int64( float128 a STATUS_PARAM ) |
| 4478 |
{
|
| 4479 |
flag aSign; |
| 4480 |
int32 aExp, shiftCount; |
| 4481 |
bits64 aSig0, aSig1; |
| 4482 |
|
| 4483 |
aSig1 = extractFloat128Frac1( a ); |
| 4484 |
aSig0 = extractFloat128Frac0( a ); |
| 4485 |
aExp = extractFloat128Exp( a ); |
| 4486 |
aSign = extractFloat128Sign( a ); |
| 4487 |
if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); |
| 4488 |
shiftCount = 0x402F - aExp;
|
| 4489 |
if ( shiftCount <= 0 ) { |
| 4490 |
if ( 0x403E < aExp ) { |
| 4491 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4492 |
if ( ! aSign
|
| 4493 |
|| ( ( aExp == 0x7FFF )
|
| 4494 |
&& ( aSig1 || ( aSig0 != LIT64( 0x0001000000000000 ) ) )
|
| 4495 |
) |
| 4496 |
) {
|
| 4497 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 4498 |
} |
| 4499 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 4500 |
} |
| 4501 |
shortShift128Left( aSig0, aSig1, - shiftCount, &aSig0, &aSig1 ); |
| 4502 |
} |
| 4503 |
else {
|
| 4504 |
shift64ExtraRightJamming( aSig0, aSig1, shiftCount, &aSig0, &aSig1 ); |
| 4505 |
} |
| 4506 |
return roundAndPackInt64( aSign, aSig0, aSig1 STATUS_VAR );
|
| 4507 |
|
| 4508 |
} |
| 4509 |
|
| 4510 |
/*----------------------------------------------------------------------------
|
| 4511 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4512 |
| value `a' to the 64-bit two's complement integer format. The conversion
|
| 4513 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4514 |
| Arithmetic, except that the conversion is always rounded toward zero.
|
| 4515 |
| If `a' is a NaN, the largest positive integer is returned. Otherwise, if
|
| 4516 |
| the conversion overflows, the largest integer with the same sign as `a' is
|
| 4517 |
| returned.
|
| 4518 |
*----------------------------------------------------------------------------*/
|
| 4519 |
|
| 4520 |
int64 float128_to_int64_round_to_zero( float128 a STATUS_PARAM ) |
| 4521 |
{
|
| 4522 |
flag aSign; |
| 4523 |
int32 aExp, shiftCount; |
| 4524 |
bits64 aSig0, aSig1; |
| 4525 |
int64 z; |
| 4526 |
|
| 4527 |
aSig1 = extractFloat128Frac1( a ); |
| 4528 |
aSig0 = extractFloat128Frac0( a ); |
| 4529 |
aExp = extractFloat128Exp( a ); |
| 4530 |
aSign = extractFloat128Sign( a ); |
| 4531 |
if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); |
| 4532 |
shiftCount = aExp - 0x402F;
|
| 4533 |
if ( 0 < shiftCount ) { |
| 4534 |
if ( 0x403E <= aExp ) { |
| 4535 |
aSig0 &= LIT64( 0x0000FFFFFFFFFFFF );
|
| 4536 |
if ( ( a.high == LIT64( 0xC03E000000000000 ) ) |
| 4537 |
&& ( aSig1 < LIT64( 0x0002000000000000 ) ) ) {
|
| 4538 |
if ( aSig1 ) STATUS(float_exception_flags) |= float_flag_inexact;
|
| 4539 |
} |
| 4540 |
else {
|
| 4541 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4542 |
if ( ! aSign || ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) ) { |
| 4543 |
return LIT64( 0x7FFFFFFFFFFFFFFF ); |
| 4544 |
} |
| 4545 |
} |
| 4546 |
return (sbits64) LIT64( 0x8000000000000000 ); |
| 4547 |
} |
| 4548 |
z = ( aSig0<<shiftCount ) | ( aSig1>>( ( - shiftCount ) & 63 ) );
|
| 4549 |
if ( (bits64) ( aSig1<<shiftCount ) ) {
|
| 4550 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4551 |
} |
| 4552 |
} |
| 4553 |
else {
|
| 4554 |
if ( aExp < 0x3FFF ) { |
| 4555 |
if ( aExp | aSig0 | aSig1 ) {
|
| 4556 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4557 |
} |
| 4558 |
return 0; |
| 4559 |
} |
| 4560 |
z = aSig0>>( - shiftCount ); |
| 4561 |
if ( aSig1
|
| 4562 |
|| ( shiftCount && (bits64) ( aSig0<<( shiftCount & 63 ) ) ) ) {
|
| 4563 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4564 |
} |
| 4565 |
} |
| 4566 |
if ( aSign ) z = - z;
|
| 4567 |
return z;
|
| 4568 |
|
| 4569 |
} |
| 4570 |
|
| 4571 |
/*----------------------------------------------------------------------------
|
| 4572 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4573 |
| value `a' to the single-precision floating-point format. The conversion
|
| 4574 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4575 |
| Arithmetic.
|
| 4576 |
*----------------------------------------------------------------------------*/
|
| 4577 |
|
| 4578 |
float32 float128_to_float32( float128 a STATUS_PARAM ) |
| 4579 |
{
|
| 4580 |
flag aSign; |
| 4581 |
int32 aExp; |
| 4582 |
bits64 aSig0, aSig1; |
| 4583 |
bits32 zSig; |
| 4584 |
|
| 4585 |
aSig1 = extractFloat128Frac1( a ); |
| 4586 |
aSig0 = extractFloat128Frac0( a ); |
| 4587 |
aExp = extractFloat128Exp( a ); |
| 4588 |
aSign = extractFloat128Sign( a ); |
| 4589 |
if ( aExp == 0x7FFF ) { |
| 4590 |
if ( aSig0 | aSig1 ) {
|
| 4591 |
return commonNaNToFloat32( float128ToCommonNaN( a STATUS_VAR ) );
|
| 4592 |
} |
| 4593 |
return packFloat32( aSign, 0xFF, 0 ); |
| 4594 |
} |
| 4595 |
aSig0 |= ( aSig1 != 0 );
|
| 4596 |
shift64RightJamming( aSig0, 18, &aSig0 );
|
| 4597 |
zSig = aSig0; |
| 4598 |
if ( aExp || zSig ) {
|
| 4599 |
zSig |= 0x40000000;
|
| 4600 |
aExp -= 0x3F81;
|
| 4601 |
} |
| 4602 |
return roundAndPackFloat32( aSign, aExp, zSig STATUS_VAR );
|
| 4603 |
|
| 4604 |
} |
| 4605 |
|
| 4606 |
/*----------------------------------------------------------------------------
|
| 4607 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4608 |
| value `a' to the double-precision floating-point format. The conversion
|
| 4609 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 4610 |
| Arithmetic.
|
| 4611 |
*----------------------------------------------------------------------------*/
|
| 4612 |
|
| 4613 |
float64 float128_to_float64( float128 a STATUS_PARAM ) |
| 4614 |
{
|
| 4615 |
flag aSign; |
| 4616 |
int32 aExp; |
| 4617 |
bits64 aSig0, aSig1; |
| 4618 |
|
| 4619 |
aSig1 = extractFloat128Frac1( a ); |
| 4620 |
aSig0 = extractFloat128Frac0( a ); |
| 4621 |
aExp = extractFloat128Exp( a ); |
| 4622 |
aSign = extractFloat128Sign( a ); |
| 4623 |
if ( aExp == 0x7FFF ) { |
| 4624 |
if ( aSig0 | aSig1 ) {
|
| 4625 |
return commonNaNToFloat64( float128ToCommonNaN( a STATUS_VAR ) );
|
| 4626 |
} |
| 4627 |
return packFloat64( aSign, 0x7FF, 0 ); |
| 4628 |
} |
| 4629 |
shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 );
|
| 4630 |
aSig0 |= ( aSig1 != 0 );
|
| 4631 |
if ( aExp || aSig0 ) {
|
| 4632 |
aSig0 |= LIT64( 0x4000000000000000 );
|
| 4633 |
aExp -= 0x3C01;
|
| 4634 |
} |
| 4635 |
return roundAndPackFloat64( aSign, aExp, aSig0 STATUS_VAR );
|
| 4636 |
|
| 4637 |
} |
| 4638 |
|
| 4639 |
#ifdef FLOATX80
|
| 4640 |
|
| 4641 |
/*----------------------------------------------------------------------------
|
| 4642 |
| Returns the result of converting the quadruple-precision floating-point
|
| 4643 |
| value `a' to the extended double-precision floating-point format. The
|
| 4644 |
| conversion is performed according to the IEC/IEEE Standard for Binary
|
| 4645 |
| Floating-Point Arithmetic.
|
| 4646 |
*----------------------------------------------------------------------------*/
|
| 4647 |
|
| 4648 |
floatx80 float128_to_floatx80( float128 a STATUS_PARAM ) |
| 4649 |
{
|
| 4650 |
flag aSign; |
| 4651 |
int32 aExp; |
| 4652 |
bits64 aSig0, aSig1; |
| 4653 |
|
| 4654 |
aSig1 = extractFloat128Frac1( a ); |
| 4655 |
aSig0 = extractFloat128Frac0( a ); |
| 4656 |
aExp = extractFloat128Exp( a ); |
| 4657 |
aSign = extractFloat128Sign( a ); |
| 4658 |
if ( aExp == 0x7FFF ) { |
| 4659 |
if ( aSig0 | aSig1 ) {
|
| 4660 |
return commonNaNToFloatx80( float128ToCommonNaN( a STATUS_VAR ) );
|
| 4661 |
} |
| 4662 |
return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); |
| 4663 |
} |
| 4664 |
if ( aExp == 0 ) { |
| 4665 |
if ( ( aSig0 | aSig1 ) == 0 ) return packFloatx80( aSign, 0, 0 ); |
| 4666 |
normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); |
| 4667 |
} |
| 4668 |
else {
|
| 4669 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 4670 |
} |
| 4671 |
shortShift128Left( aSig0, aSig1, 15, &aSig0, &aSig1 );
|
| 4672 |
return roundAndPackFloatx80( 80, aSign, aExp, aSig0, aSig1 STATUS_VAR ); |
| 4673 |
|
| 4674 |
} |
| 4675 |
|
| 4676 |
#endif
|
| 4677 |
|
| 4678 |
/*----------------------------------------------------------------------------
|
| 4679 |
| Rounds the quadruple-precision floating-point value `a' to an integer, and
|
| 4680 |
| returns the result as a quadruple-precision floating-point value. The
|
| 4681 |
| operation is performed according to the IEC/IEEE Standard for Binary
|
| 4682 |
| Floating-Point Arithmetic.
|
| 4683 |
*----------------------------------------------------------------------------*/
|
| 4684 |
|
| 4685 |
float128 float128_round_to_int( float128 a STATUS_PARAM ) |
| 4686 |
{
|
| 4687 |
flag aSign; |
| 4688 |
int32 aExp; |
| 4689 |
bits64 lastBitMask, roundBitsMask; |
| 4690 |
int8 roundingMode; |
| 4691 |
float128 z; |
| 4692 |
|
| 4693 |
aExp = extractFloat128Exp( a ); |
| 4694 |
if ( 0x402F <= aExp ) { |
| 4695 |
if ( 0x406F <= aExp ) { |
| 4696 |
if ( ( aExp == 0x7FFF ) |
| 4697 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) |
| 4698 |
) {
|
| 4699 |
return propagateFloat128NaN( a, a STATUS_VAR );
|
| 4700 |
} |
| 4701 |
return a;
|
| 4702 |
} |
| 4703 |
lastBitMask = 1;
|
| 4704 |
lastBitMask = ( lastBitMask<<( 0x406E - aExp ) )<<1; |
| 4705 |
roundBitsMask = lastBitMask - 1;
|
| 4706 |
z = a; |
| 4707 |
roundingMode = STATUS(float_rounding_mode); |
| 4708 |
if ( roundingMode == float_round_nearest_even ) {
|
| 4709 |
if ( lastBitMask ) {
|
| 4710 |
add128( z.high, z.low, 0, lastBitMask>>1, &z.high, &z.low ); |
| 4711 |
if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; |
| 4712 |
} |
| 4713 |
else {
|
| 4714 |
if ( (sbits64) z.low < 0 ) { |
| 4715 |
++z.high; |
| 4716 |
if ( (bits64) ( z.low<<1 ) == 0 ) z.high &= ~1; |
| 4717 |
} |
| 4718 |
} |
| 4719 |
} |
| 4720 |
else if ( roundingMode != float_round_to_zero ) { |
| 4721 |
if ( extractFloat128Sign( z )
|
| 4722 |
^ ( roundingMode == float_round_up ) ) {
|
| 4723 |
add128( z.high, z.low, 0, roundBitsMask, &z.high, &z.low );
|
| 4724 |
} |
| 4725 |
} |
| 4726 |
z.low &= ~ roundBitsMask; |
| 4727 |
} |
| 4728 |
else {
|
| 4729 |
if ( aExp < 0x3FFF ) { |
| 4730 |
if ( ( ( (bits64) ( a.high<<1 ) ) | a.low ) == 0 ) return a; |
| 4731 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4732 |
aSign = extractFloat128Sign( a ); |
| 4733 |
switch ( STATUS(float_rounding_mode) ) {
|
| 4734 |
case float_round_nearest_even:
|
| 4735 |
if ( ( aExp == 0x3FFE ) |
| 4736 |
&& ( extractFloat128Frac0( a ) |
| 4737 |
| extractFloat128Frac1( a ) ) |
| 4738 |
) {
|
| 4739 |
return packFloat128( aSign, 0x3FFF, 0, 0 ); |
| 4740 |
} |
| 4741 |
break;
|
| 4742 |
case float_round_down:
|
| 4743 |
return
|
| 4744 |
aSign ? packFloat128( 1, 0x3FFF, 0, 0 ) |
| 4745 |
: packFloat128( 0, 0, 0, 0 ); |
| 4746 |
case float_round_up:
|
| 4747 |
return
|
| 4748 |
aSign ? packFloat128( 1, 0, 0, 0 ) |
| 4749 |
: packFloat128( 0, 0x3FFF, 0, 0 ); |
| 4750 |
} |
| 4751 |
return packFloat128( aSign, 0, 0, 0 ); |
| 4752 |
} |
| 4753 |
lastBitMask = 1;
|
| 4754 |
lastBitMask <<= 0x402F - aExp;
|
| 4755 |
roundBitsMask = lastBitMask - 1;
|
| 4756 |
z.low = 0;
|
| 4757 |
z.high = a.high; |
| 4758 |
roundingMode = STATUS(float_rounding_mode); |
| 4759 |
if ( roundingMode == float_round_nearest_even ) {
|
| 4760 |
z.high += lastBitMask>>1;
|
| 4761 |
if ( ( ( z.high & roundBitsMask ) | a.low ) == 0 ) { |
| 4762 |
z.high &= ~ lastBitMask; |
| 4763 |
} |
| 4764 |
} |
| 4765 |
else if ( roundingMode != float_round_to_zero ) { |
| 4766 |
if ( extractFloat128Sign( z )
|
| 4767 |
^ ( roundingMode == float_round_up ) ) {
|
| 4768 |
z.high |= ( a.low != 0 );
|
| 4769 |
z.high += roundBitsMask; |
| 4770 |
} |
| 4771 |
} |
| 4772 |
z.high &= ~ roundBitsMask; |
| 4773 |
} |
| 4774 |
if ( ( z.low != a.low ) || ( z.high != a.high ) ) {
|
| 4775 |
STATUS(float_exception_flags) |= float_flag_inexact; |
| 4776 |
} |
| 4777 |
return z;
|
| 4778 |
|
| 4779 |
} |
| 4780 |
|
| 4781 |
/*----------------------------------------------------------------------------
|
| 4782 |
| Returns the result of adding the absolute values of the quadruple-precision
|
| 4783 |
| floating-point values `a' and `b'. If `zSign' is 1, the sum is negated
|
| 4784 |
| before being returned. `zSign' is ignored if the result is a NaN.
|
| 4785 |
| The addition is performed according to the IEC/IEEE Standard for Binary
|
| 4786 |
| Floating-Point Arithmetic.
|
| 4787 |
*----------------------------------------------------------------------------*/
|
| 4788 |
|
| 4789 |
static float128 addFloat128Sigs( float128 a, float128 b, flag zSign STATUS_PARAM)
|
| 4790 |
{
|
| 4791 |
int32 aExp, bExp, zExp; |
| 4792 |
bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; |
| 4793 |
int32 expDiff; |
| 4794 |
|
| 4795 |
aSig1 = extractFloat128Frac1( a ); |
| 4796 |
aSig0 = extractFloat128Frac0( a ); |
| 4797 |
aExp = extractFloat128Exp( a ); |
| 4798 |
bSig1 = extractFloat128Frac1( b ); |
| 4799 |
bSig0 = extractFloat128Frac0( b ); |
| 4800 |
bExp = extractFloat128Exp( b ); |
| 4801 |
expDiff = aExp - bExp; |
| 4802 |
if ( 0 < expDiff ) { |
| 4803 |
if ( aExp == 0x7FFF ) { |
| 4804 |
if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 4805 |
return a;
|
| 4806 |
} |
| 4807 |
if ( bExp == 0 ) { |
| 4808 |
--expDiff; |
| 4809 |
} |
| 4810 |
else {
|
| 4811 |
bSig0 |= LIT64( 0x0001000000000000 );
|
| 4812 |
} |
| 4813 |
shift128ExtraRightJamming( |
| 4814 |
bSig0, bSig1, 0, expDiff, &bSig0, &bSig1, &zSig2 );
|
| 4815 |
zExp = aExp; |
| 4816 |
} |
| 4817 |
else if ( expDiff < 0 ) { |
| 4818 |
if ( bExp == 0x7FFF ) { |
| 4819 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 4820 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 4821 |
} |
| 4822 |
if ( aExp == 0 ) { |
| 4823 |
++expDiff; |
| 4824 |
} |
| 4825 |
else {
|
| 4826 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 4827 |
} |
| 4828 |
shift128ExtraRightJamming( |
| 4829 |
aSig0, aSig1, 0, - expDiff, &aSig0, &aSig1, &zSig2 );
|
| 4830 |
zExp = bExp; |
| 4831 |
} |
| 4832 |
else {
|
| 4833 |
if ( aExp == 0x7FFF ) { |
| 4834 |
if ( aSig0 | aSig1 | bSig0 | bSig1 ) {
|
| 4835 |
return propagateFloat128NaN( a, b STATUS_VAR );
|
| 4836 |
} |
| 4837 |
return a;
|
| 4838 |
} |
| 4839 |
add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); |
| 4840 |
if ( aExp == 0 ) { |
| 4841 |
if ( STATUS(flush_to_zero) ) return packFloat128( zSign, 0, 0, 0 ); |
| 4842 |
return packFloat128( zSign, 0, zSig0, zSig1 ); |
| 4843 |
} |
| 4844 |
zSig2 = 0;
|
| 4845 |
zSig0 |= LIT64( 0x0002000000000000 );
|
| 4846 |
zExp = aExp; |
| 4847 |
goto shiftRight1;
|
| 4848 |
} |
| 4849 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 4850 |
add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); |
| 4851 |
--zExp; |
| 4852 |
if ( zSig0 < LIT64( 0x0002000000000000 ) ) goto roundAndPack; |
| 4853 |
++zExp; |
| 4854 |
shiftRight1:
|
| 4855 |
shift128ExtraRightJamming( |
| 4856 |
zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 );
|
| 4857 |
roundAndPack:
|
| 4858 |
return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 STATUS_VAR );
|
| 4859 |
|
| 4860 |
} |
| 4861 |
|
| 4862 |
/*----------------------------------------------------------------------------
|
| 4863 |
| Returns the result of subtracting the absolute values of the quadruple-
|
| 4864 |
| precision floating-point values `a' and `b'. If `zSign' is 1, the
|
| 4865 |
| difference is negated before being returned. `zSign' is ignored if the
|
| 4866 |
| result is a NaN. The subtraction is performed according to the IEC/IEEE
|
| 4867 |
| Standard for Binary Floating-Point Arithmetic.
|
| 4868 |
*----------------------------------------------------------------------------*/
|
| 4869 |
|
| 4870 |
static float128 subFloat128Sigs( float128 a, float128 b, flag zSign STATUS_PARAM)
|
| 4871 |
{
|
| 4872 |
int32 aExp, bExp, zExp; |
| 4873 |
bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1; |
| 4874 |
int32 expDiff; |
| 4875 |
float128 z; |
| 4876 |
|
| 4877 |
aSig1 = extractFloat128Frac1( a ); |
| 4878 |
aSig0 = extractFloat128Frac0( a ); |
| 4879 |
aExp = extractFloat128Exp( a ); |
| 4880 |
bSig1 = extractFloat128Frac1( b ); |
| 4881 |
bSig0 = extractFloat128Frac0( b ); |
| 4882 |
bExp = extractFloat128Exp( b ); |
| 4883 |
expDiff = aExp - bExp; |
| 4884 |
shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 );
|
| 4885 |
shortShift128Left( bSig0, bSig1, 14, &bSig0, &bSig1 );
|
| 4886 |
if ( 0 < expDiff ) goto aExpBigger; |
| 4887 |
if ( expDiff < 0 ) goto bExpBigger; |
| 4888 |
if ( aExp == 0x7FFF ) { |
| 4889 |
if ( aSig0 | aSig1 | bSig0 | bSig1 ) {
|
| 4890 |
return propagateFloat128NaN( a, b STATUS_VAR );
|
| 4891 |
} |
| 4892 |
float_raise( float_flag_invalid STATUS_VAR); |
| 4893 |
z.low = float128_default_nan_low; |
| 4894 |
z.high = float128_default_nan_high; |
| 4895 |
return z;
|
| 4896 |
} |
| 4897 |
if ( aExp == 0 ) { |
| 4898 |
aExp = 1;
|
| 4899 |
bExp = 1;
|
| 4900 |
} |
| 4901 |
if ( bSig0 < aSig0 ) goto aBigger; |
| 4902 |
if ( aSig0 < bSig0 ) goto bBigger; |
| 4903 |
if ( bSig1 < aSig1 ) goto aBigger; |
| 4904 |
if ( aSig1 < bSig1 ) goto bBigger; |
| 4905 |
return packFloat128( STATUS(float_rounding_mode) == float_round_down, 0, 0, 0 ); |
| 4906 |
bExpBigger:
|
| 4907 |
if ( bExp == 0x7FFF ) { |
| 4908 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 4909 |
return packFloat128( zSign ^ 1, 0x7FFF, 0, 0 ); |
| 4910 |
} |
| 4911 |
if ( aExp == 0 ) { |
| 4912 |
++expDiff; |
| 4913 |
} |
| 4914 |
else {
|
| 4915 |
aSig0 |= LIT64( 0x4000000000000000 );
|
| 4916 |
} |
| 4917 |
shift128RightJamming( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); |
| 4918 |
bSig0 |= LIT64( 0x4000000000000000 );
|
| 4919 |
bBigger:
|
| 4920 |
sub128( bSig0, bSig1, aSig0, aSig1, &zSig0, &zSig1 ); |
| 4921 |
zExp = bExp; |
| 4922 |
zSign ^= 1;
|
| 4923 |
goto normalizeRoundAndPack;
|
| 4924 |
aExpBigger:
|
| 4925 |
if ( aExp == 0x7FFF ) { |
| 4926 |
if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 4927 |
return a;
|
| 4928 |
} |
| 4929 |
if ( bExp == 0 ) { |
| 4930 |
--expDiff; |
| 4931 |
} |
| 4932 |
else {
|
| 4933 |
bSig0 |= LIT64( 0x4000000000000000 );
|
| 4934 |
} |
| 4935 |
shift128RightJamming( bSig0, bSig1, expDiff, &bSig0, &bSig1 ); |
| 4936 |
aSig0 |= LIT64( 0x4000000000000000 );
|
| 4937 |
aBigger:
|
| 4938 |
sub128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); |
| 4939 |
zExp = aExp; |
| 4940 |
normalizeRoundAndPack:
|
| 4941 |
--zExp; |
| 4942 |
return normalizeRoundAndPackFloat128( zSign, zExp - 14, zSig0, zSig1 STATUS_VAR ); |
| 4943 |
|
| 4944 |
} |
| 4945 |
|
| 4946 |
/*----------------------------------------------------------------------------
|
| 4947 |
| Returns the result of adding the quadruple-precision floating-point values
|
| 4948 |
| `a' and `b'. The operation is performed according to the IEC/IEEE Standard
|
| 4949 |
| for Binary Floating-Point Arithmetic.
|
| 4950 |
*----------------------------------------------------------------------------*/
|
| 4951 |
|
| 4952 |
float128 float128_add( float128 a, float128 b STATUS_PARAM ) |
| 4953 |
{
|
| 4954 |
flag aSign, bSign; |
| 4955 |
|
| 4956 |
aSign = extractFloat128Sign( a ); |
| 4957 |
bSign = extractFloat128Sign( b ); |
| 4958 |
if ( aSign == bSign ) {
|
| 4959 |
return addFloat128Sigs( a, b, aSign STATUS_VAR );
|
| 4960 |
} |
| 4961 |
else {
|
| 4962 |
return subFloat128Sigs( a, b, aSign STATUS_VAR );
|
| 4963 |
} |
| 4964 |
|
| 4965 |
} |
| 4966 |
|
| 4967 |
/*----------------------------------------------------------------------------
|
| 4968 |
| Returns the result of subtracting the quadruple-precision floating-point
|
| 4969 |
| values `a' and `b'. The operation is performed according to the IEC/IEEE
|
| 4970 |
| Standard for Binary Floating-Point Arithmetic.
|
| 4971 |
*----------------------------------------------------------------------------*/
|
| 4972 |
|
| 4973 |
float128 float128_sub( float128 a, float128 b STATUS_PARAM ) |
| 4974 |
{
|
| 4975 |
flag aSign, bSign; |
| 4976 |
|
| 4977 |
aSign = extractFloat128Sign( a ); |
| 4978 |
bSign = extractFloat128Sign( b ); |
| 4979 |
if ( aSign == bSign ) {
|
| 4980 |
return subFloat128Sigs( a, b, aSign STATUS_VAR );
|
| 4981 |
} |
| 4982 |
else {
|
| 4983 |
return addFloat128Sigs( a, b, aSign STATUS_VAR );
|
| 4984 |
} |
| 4985 |
|
| 4986 |
} |
| 4987 |
|
| 4988 |
/*----------------------------------------------------------------------------
|
| 4989 |
| Returns the result of multiplying the quadruple-precision floating-point
|
| 4990 |
| values `a' and `b'. The operation is performed according to the IEC/IEEE
|
| 4991 |
| Standard for Binary Floating-Point Arithmetic.
|
| 4992 |
*----------------------------------------------------------------------------*/
|
| 4993 |
|
| 4994 |
float128 float128_mul( float128 a, float128 b STATUS_PARAM ) |
| 4995 |
{
|
| 4996 |
flag aSign, bSign, zSign; |
| 4997 |
int32 aExp, bExp, zExp; |
| 4998 |
bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2, zSig3; |
| 4999 |
float128 z; |
| 5000 |
|
| 5001 |
aSig1 = extractFloat128Frac1( a ); |
| 5002 |
aSig0 = extractFloat128Frac0( a ); |
| 5003 |
aExp = extractFloat128Exp( a ); |
| 5004 |
aSign = extractFloat128Sign( a ); |
| 5005 |
bSig1 = extractFloat128Frac1( b ); |
| 5006 |
bSig0 = extractFloat128Frac0( b ); |
| 5007 |
bExp = extractFloat128Exp( b ); |
| 5008 |
bSign = extractFloat128Sign( b ); |
| 5009 |
zSign = aSign ^ bSign; |
| 5010 |
if ( aExp == 0x7FFF ) { |
| 5011 |
if ( ( aSig0 | aSig1 )
|
| 5012 |
|| ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) {
|
| 5013 |
return propagateFloat128NaN( a, b STATUS_VAR );
|
| 5014 |
} |
| 5015 |
if ( ( bExp | bSig0 | bSig1 ) == 0 ) goto invalid; |
| 5016 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 5017 |
} |
| 5018 |
if ( bExp == 0x7FFF ) { |
| 5019 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 5020 |
if ( ( aExp | aSig0 | aSig1 ) == 0 ) { |
| 5021 |
invalid:
|
| 5022 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5023 |
z.low = float128_default_nan_low; |
| 5024 |
z.high = float128_default_nan_high; |
| 5025 |
return z;
|
| 5026 |
} |
| 5027 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 5028 |
} |
| 5029 |
if ( aExp == 0 ) { |
| 5030 |
if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); |
| 5031 |
normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); |
| 5032 |
} |
| 5033 |
if ( bExp == 0 ) { |
| 5034 |
if ( ( bSig0 | bSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); |
| 5035 |
normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); |
| 5036 |
} |
| 5037 |
zExp = aExp + bExp - 0x4000;
|
| 5038 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 5039 |
shortShift128Left( bSig0, bSig1, 16, &bSig0, &bSig1 );
|
| 5040 |
mul128To256( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1, &zSig2, &zSig3 ); |
| 5041 |
add128( zSig0, zSig1, aSig0, aSig1, &zSig0, &zSig1 ); |
| 5042 |
zSig2 |= ( zSig3 != 0 );
|
| 5043 |
if ( LIT64( 0x0002000000000000 ) <= zSig0 ) { |
| 5044 |
shift128ExtraRightJamming( |
| 5045 |
zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 );
|
| 5046 |
++zExp; |
| 5047 |
} |
| 5048 |
return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 STATUS_VAR );
|
| 5049 |
|
| 5050 |
} |
| 5051 |
|
| 5052 |
/*----------------------------------------------------------------------------
|
| 5053 |
| Returns the result of dividing the quadruple-precision floating-point value
|
| 5054 |
| `a' by the corresponding value `b'. The operation is performed according to
|
| 5055 |
| the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5056 |
*----------------------------------------------------------------------------*/
|
| 5057 |
|
| 5058 |
float128 float128_div( float128 a, float128 b STATUS_PARAM ) |
| 5059 |
{
|
| 5060 |
flag aSign, bSign, zSign; |
| 5061 |
int32 aExp, bExp, zExp; |
| 5062 |
bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; |
| 5063 |
bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; |
| 5064 |
float128 z; |
| 5065 |
|
| 5066 |
aSig1 = extractFloat128Frac1( a ); |
| 5067 |
aSig0 = extractFloat128Frac0( a ); |
| 5068 |
aExp = extractFloat128Exp( a ); |
| 5069 |
aSign = extractFloat128Sign( a ); |
| 5070 |
bSig1 = extractFloat128Frac1( b ); |
| 5071 |
bSig0 = extractFloat128Frac0( b ); |
| 5072 |
bExp = extractFloat128Exp( b ); |
| 5073 |
bSign = extractFloat128Sign( b ); |
| 5074 |
zSign = aSign ^ bSign; |
| 5075 |
if ( aExp == 0x7FFF ) { |
| 5076 |
if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 5077 |
if ( bExp == 0x7FFF ) { |
| 5078 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 5079 |
goto invalid;
|
| 5080 |
} |
| 5081 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 5082 |
} |
| 5083 |
if ( bExp == 0x7FFF ) { |
| 5084 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 5085 |
return packFloat128( zSign, 0, 0, 0 ); |
| 5086 |
} |
| 5087 |
if ( bExp == 0 ) { |
| 5088 |
if ( ( bSig0 | bSig1 ) == 0 ) { |
| 5089 |
if ( ( aExp | aSig0 | aSig1 ) == 0 ) { |
| 5090 |
invalid:
|
| 5091 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5092 |
z.low = float128_default_nan_low; |
| 5093 |
z.high = float128_default_nan_high; |
| 5094 |
return z;
|
| 5095 |
} |
| 5096 |
float_raise( float_flag_divbyzero STATUS_VAR); |
| 5097 |
return packFloat128( zSign, 0x7FFF, 0, 0 ); |
| 5098 |
} |
| 5099 |
normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); |
| 5100 |
} |
| 5101 |
if ( aExp == 0 ) { |
| 5102 |
if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); |
| 5103 |
normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); |
| 5104 |
} |
| 5105 |
zExp = aExp - bExp + 0x3FFD;
|
| 5106 |
shortShift128Left( |
| 5107 |
aSig0 | LIT64( 0x0001000000000000 ), aSig1, 15, &aSig0, &aSig1 ); |
| 5108 |
shortShift128Left( |
| 5109 |
bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); |
| 5110 |
if ( le128( bSig0, bSig1, aSig0, aSig1 ) ) {
|
| 5111 |
shift128Right( aSig0, aSig1, 1, &aSig0, &aSig1 );
|
| 5112 |
++zExp; |
| 5113 |
} |
| 5114 |
zSig0 = estimateDiv128To64( aSig0, aSig1, bSig0 ); |
| 5115 |
mul128By64To192( bSig0, bSig1, zSig0, &term0, &term1, &term2 ); |
| 5116 |
sub192( aSig0, aSig1, 0, term0, term1, term2, &rem0, &rem1, &rem2 );
|
| 5117 |
while ( (sbits64) rem0 < 0 ) { |
| 5118 |
--zSig0; |
| 5119 |
add192( rem0, rem1, rem2, 0, bSig0, bSig1, &rem0, &rem1, &rem2 );
|
| 5120 |
} |
| 5121 |
zSig1 = estimateDiv128To64( rem1, rem2, bSig0 ); |
| 5122 |
if ( ( zSig1 & 0x3FFF ) <= 4 ) { |
| 5123 |
mul128By64To192( bSig0, bSig1, zSig1, &term1, &term2, &term3 ); |
| 5124 |
sub192( rem1, rem2, 0, term1, term2, term3, &rem1, &rem2, &rem3 );
|
| 5125 |
while ( (sbits64) rem1 < 0 ) { |
| 5126 |
--zSig1; |
| 5127 |
add192( rem1, rem2, rem3, 0, bSig0, bSig1, &rem1, &rem2, &rem3 );
|
| 5128 |
} |
| 5129 |
zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 );
|
| 5130 |
} |
| 5131 |
shift128ExtraRightJamming( zSig0, zSig1, 0, 15, &zSig0, &zSig1, &zSig2 ); |
| 5132 |
return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 STATUS_VAR );
|
| 5133 |
|
| 5134 |
} |
| 5135 |
|
| 5136 |
/*----------------------------------------------------------------------------
|
| 5137 |
| Returns the remainder of the quadruple-precision floating-point value `a'
|
| 5138 |
| with respect to the corresponding value `b'. The operation is performed
|
| 5139 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5140 |
*----------------------------------------------------------------------------*/
|
| 5141 |
|
| 5142 |
float128 float128_rem( float128 a, float128 b STATUS_PARAM ) |
| 5143 |
{
|
| 5144 |
flag aSign, zSign; |
| 5145 |
int32 aExp, bExp, expDiff; |
| 5146 |
bits64 aSig0, aSig1, bSig0, bSig1, q, term0, term1, term2; |
| 5147 |
bits64 allZero, alternateASig0, alternateASig1, sigMean1; |
| 5148 |
sbits64 sigMean0; |
| 5149 |
float128 z; |
| 5150 |
|
| 5151 |
aSig1 = extractFloat128Frac1( a ); |
| 5152 |
aSig0 = extractFloat128Frac0( a ); |
| 5153 |
aExp = extractFloat128Exp( a ); |
| 5154 |
aSign = extractFloat128Sign( a ); |
| 5155 |
bSig1 = extractFloat128Frac1( b ); |
| 5156 |
bSig0 = extractFloat128Frac0( b ); |
| 5157 |
bExp = extractFloat128Exp( b ); |
| 5158 |
if ( aExp == 0x7FFF ) { |
| 5159 |
if ( ( aSig0 | aSig1 )
|
| 5160 |
|| ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) {
|
| 5161 |
return propagateFloat128NaN( a, b STATUS_VAR );
|
| 5162 |
} |
| 5163 |
goto invalid;
|
| 5164 |
} |
| 5165 |
if ( bExp == 0x7FFF ) { |
| 5166 |
if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b STATUS_VAR ); |
| 5167 |
return a;
|
| 5168 |
} |
| 5169 |
if ( bExp == 0 ) { |
| 5170 |
if ( ( bSig0 | bSig1 ) == 0 ) { |
| 5171 |
invalid:
|
| 5172 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5173 |
z.low = float128_default_nan_low; |
| 5174 |
z.high = float128_default_nan_high; |
| 5175 |
return z;
|
| 5176 |
} |
| 5177 |
normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); |
| 5178 |
} |
| 5179 |
if ( aExp == 0 ) { |
| 5180 |
if ( ( aSig0 | aSig1 ) == 0 ) return a; |
| 5181 |
normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); |
| 5182 |
} |
| 5183 |
expDiff = aExp - bExp; |
| 5184 |
if ( expDiff < -1 ) return a; |
| 5185 |
shortShift128Left( |
| 5186 |
aSig0 | LIT64( 0x0001000000000000 ),
|
| 5187 |
aSig1, |
| 5188 |
15 - ( expDiff < 0 ), |
| 5189 |
&aSig0, |
| 5190 |
&aSig1 |
| 5191 |
); |
| 5192 |
shortShift128Left( |
| 5193 |
bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); |
| 5194 |
q = le128( bSig0, bSig1, aSig0, aSig1 ); |
| 5195 |
if ( q ) sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 );
|
| 5196 |
expDiff -= 64;
|
| 5197 |
while ( 0 < expDiff ) { |
| 5198 |
q = estimateDiv128To64( aSig0, aSig1, bSig0 ); |
| 5199 |
q = ( 4 < q ) ? q - 4 : 0; |
| 5200 |
mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); |
| 5201 |
shortShift192Left( term0, term1, term2, 61, &term1, &term2, &allZero );
|
| 5202 |
shortShift128Left( aSig0, aSig1, 61, &aSig0, &allZero );
|
| 5203 |
sub128( aSig0, 0, term1, term2, &aSig0, &aSig1 );
|
| 5204 |
expDiff -= 61;
|
| 5205 |
} |
| 5206 |
if ( -64 < expDiff ) { |
| 5207 |
q = estimateDiv128To64( aSig0, aSig1, bSig0 ); |
| 5208 |
q = ( 4 < q ) ? q - 4 : 0; |
| 5209 |
q >>= - expDiff; |
| 5210 |
shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 );
|
| 5211 |
expDiff += 52;
|
| 5212 |
if ( expDiff < 0 ) { |
| 5213 |
shift128Right( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); |
| 5214 |
} |
| 5215 |
else {
|
| 5216 |
shortShift128Left( aSig0, aSig1, expDiff, &aSig0, &aSig1 ); |
| 5217 |
} |
| 5218 |
mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); |
| 5219 |
sub128( aSig0, aSig1, term1, term2, &aSig0, &aSig1 ); |
| 5220 |
} |
| 5221 |
else {
|
| 5222 |
shift128Right( aSig0, aSig1, 12, &aSig0, &aSig1 );
|
| 5223 |
shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 );
|
| 5224 |
} |
| 5225 |
do {
|
| 5226 |
alternateASig0 = aSig0; |
| 5227 |
alternateASig1 = aSig1; |
| 5228 |
++q; |
| 5229 |
sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 ); |
| 5230 |
} while ( 0 <= (sbits64) aSig0 ); |
| 5231 |
add128( |
| 5232 |
aSig0, aSig1, alternateASig0, alternateASig1, (bits64 *)&sigMean0, &sigMean1 ); |
| 5233 |
if ( ( sigMean0 < 0 ) |
| 5234 |
|| ( ( ( sigMean0 | sigMean1 ) == 0 ) && ( q & 1 ) ) ) { |
| 5235 |
aSig0 = alternateASig0; |
| 5236 |
aSig1 = alternateASig1; |
| 5237 |
} |
| 5238 |
zSign = ( (sbits64) aSig0 < 0 );
|
| 5239 |
if ( zSign ) sub128( 0, 0, aSig0, aSig1, &aSig0, &aSig1 ); |
| 5240 |
return
|
| 5241 |
normalizeRoundAndPackFloat128( aSign ^ zSign, bExp - 4, aSig0, aSig1 STATUS_VAR );
|
| 5242 |
|
| 5243 |
} |
| 5244 |
|
| 5245 |
/*----------------------------------------------------------------------------
|
| 5246 |
| Returns the square root of the quadruple-precision floating-point value `a'.
|
| 5247 |
| The operation is performed according to the IEC/IEEE Standard for Binary
|
| 5248 |
| Floating-Point Arithmetic.
|
| 5249 |
*----------------------------------------------------------------------------*/
|
| 5250 |
|
| 5251 |
float128 float128_sqrt( float128 a STATUS_PARAM ) |
| 5252 |
{
|
| 5253 |
flag aSign; |
| 5254 |
int32 aExp, zExp; |
| 5255 |
bits64 aSig0, aSig1, zSig0, zSig1, zSig2, doubleZSig0; |
| 5256 |
bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; |
| 5257 |
float128 z; |
| 5258 |
|
| 5259 |
aSig1 = extractFloat128Frac1( a ); |
| 5260 |
aSig0 = extractFloat128Frac0( a ); |
| 5261 |
aExp = extractFloat128Exp( a ); |
| 5262 |
aSign = extractFloat128Sign( a ); |
| 5263 |
if ( aExp == 0x7FFF ) { |
| 5264 |
if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, a STATUS_VAR ); |
| 5265 |
if ( ! aSign ) return a; |
| 5266 |
goto invalid;
|
| 5267 |
} |
| 5268 |
if ( aSign ) {
|
| 5269 |
if ( ( aExp | aSig0 | aSig1 ) == 0 ) return a; |
| 5270 |
invalid:
|
| 5271 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5272 |
z.low = float128_default_nan_low; |
| 5273 |
z.high = float128_default_nan_high; |
| 5274 |
return z;
|
| 5275 |
} |
| 5276 |
if ( aExp == 0 ) { |
| 5277 |
if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( 0, 0, 0, 0 ); |
| 5278 |
normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); |
| 5279 |
} |
| 5280 |
zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFE; |
| 5281 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 5282 |
zSig0 = estimateSqrt32( aExp, aSig0>>17 );
|
| 5283 |
shortShift128Left( aSig0, aSig1, 13 - ( aExp & 1 ), &aSig0, &aSig1 ); |
| 5284 |
zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); |
| 5285 |
doubleZSig0 = zSig0<<1;
|
| 5286 |
mul64To128( zSig0, zSig0, &term0, &term1 ); |
| 5287 |
sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); |
| 5288 |
while ( (sbits64) rem0 < 0 ) { |
| 5289 |
--zSig0; |
| 5290 |
doubleZSig0 -= 2;
|
| 5291 |
add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); |
| 5292 |
} |
| 5293 |
zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 );
|
| 5294 |
if ( ( zSig1 & 0x1FFF ) <= 5 ) { |
| 5295 |
if ( zSig1 == 0 ) zSig1 = 1; |
| 5296 |
mul64To128( doubleZSig0, zSig1, &term1, &term2 ); |
| 5297 |
sub128( rem1, 0, term1, term2, &rem1, &rem2 );
|
| 5298 |
mul64To128( zSig1, zSig1, &term2, &term3 ); |
| 5299 |
sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); |
| 5300 |
while ( (sbits64) rem1 < 0 ) { |
| 5301 |
--zSig1; |
| 5302 |
shortShift128Left( 0, zSig1, 1, &term2, &term3 ); |
| 5303 |
term3 |= 1;
|
| 5304 |
term2 |= doubleZSig0; |
| 5305 |
add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 );
|
| 5306 |
} |
| 5307 |
zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 );
|
| 5308 |
} |
| 5309 |
shift128ExtraRightJamming( zSig0, zSig1, 0, 14, &zSig0, &zSig1, &zSig2 ); |
| 5310 |
return roundAndPackFloat128( 0, zExp, zSig0, zSig1, zSig2 STATUS_VAR ); |
| 5311 |
|
| 5312 |
} |
| 5313 |
|
| 5314 |
/*----------------------------------------------------------------------------
|
| 5315 |
| Returns 1 if the quadruple-precision floating-point value `a' is equal to
|
| 5316 |
| the corresponding value `b', and 0 otherwise. The comparison is performed
|
| 5317 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5318 |
*----------------------------------------------------------------------------*/
|
| 5319 |
|
| 5320 |
int float128_eq( float128 a, float128 b STATUS_PARAM )
|
| 5321 |
{
|
| 5322 |
|
| 5323 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5324 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5325 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5326 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5327 |
) {
|
| 5328 |
if ( float128_is_signaling_nan( a )
|
| 5329 |
|| float128_is_signaling_nan( b ) ) {
|
| 5330 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5331 |
} |
| 5332 |
return 0; |
| 5333 |
} |
| 5334 |
return
|
| 5335 |
( a.low == b.low ) |
| 5336 |
&& ( ( a.high == b.high ) |
| 5337 |
|| ( ( a.low == 0 )
|
| 5338 |
&& ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) |
| 5339 |
); |
| 5340 |
|
| 5341 |
} |
| 5342 |
|
| 5343 |
/*----------------------------------------------------------------------------
|
| 5344 |
| Returns 1 if the quadruple-precision floating-point value `a' is less than
|
| 5345 |
| or equal to the corresponding value `b', and 0 otherwise. The comparison
|
| 5346 |
| is performed according to the IEC/IEEE Standard for Binary Floating-Point
|
| 5347 |
| Arithmetic.
|
| 5348 |
*----------------------------------------------------------------------------*/
|
| 5349 |
|
| 5350 |
int float128_le( float128 a, float128 b STATUS_PARAM )
|
| 5351 |
{
|
| 5352 |
flag aSign, bSign; |
| 5353 |
|
| 5354 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5355 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5356 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5357 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5358 |
) {
|
| 5359 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5360 |
return 0; |
| 5361 |
} |
| 5362 |
aSign = extractFloat128Sign( a ); |
| 5363 |
bSign = extractFloat128Sign( b ); |
| 5364 |
if ( aSign != bSign ) {
|
| 5365 |
return
|
| 5366 |
aSign |
| 5367 |
|| ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 5368 |
== 0 );
|
| 5369 |
} |
| 5370 |
return
|
| 5371 |
aSign ? le128( b.high, b.low, a.high, a.low ) |
| 5372 |
: le128( a.high, a.low, b.high, b.low ); |
| 5373 |
|
| 5374 |
} |
| 5375 |
|
| 5376 |
/*----------------------------------------------------------------------------
|
| 5377 |
| Returns 1 if the quadruple-precision floating-point value `a' is less than
|
| 5378 |
| the corresponding value `b', and 0 otherwise. The comparison is performed
|
| 5379 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5380 |
*----------------------------------------------------------------------------*/
|
| 5381 |
|
| 5382 |
int float128_lt( float128 a, float128 b STATUS_PARAM )
|
| 5383 |
{
|
| 5384 |
flag aSign, bSign; |
| 5385 |
|
| 5386 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5387 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5388 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5389 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5390 |
) {
|
| 5391 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5392 |
return 0; |
| 5393 |
} |
| 5394 |
aSign = extractFloat128Sign( a ); |
| 5395 |
bSign = extractFloat128Sign( b ); |
| 5396 |
if ( aSign != bSign ) {
|
| 5397 |
return
|
| 5398 |
aSign |
| 5399 |
&& ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 5400 |
!= 0 );
|
| 5401 |
} |
| 5402 |
return
|
| 5403 |
aSign ? lt128( b.high, b.low, a.high, a.low ) |
| 5404 |
: lt128( a.high, a.low, b.high, b.low ); |
| 5405 |
|
| 5406 |
} |
| 5407 |
|
| 5408 |
/*----------------------------------------------------------------------------
|
| 5409 |
| Returns 1 if the quadruple-precision floating-point value `a' is equal to
|
| 5410 |
| the corresponding value `b', and 0 otherwise. The invalid exception is
|
| 5411 |
| raised if either operand is a NaN. Otherwise, the comparison is performed
|
| 5412 |
| according to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5413 |
*----------------------------------------------------------------------------*/
|
| 5414 |
|
| 5415 |
int float128_eq_signaling( float128 a, float128 b STATUS_PARAM )
|
| 5416 |
{
|
| 5417 |
|
| 5418 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5419 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5420 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5421 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5422 |
) {
|
| 5423 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5424 |
return 0; |
| 5425 |
} |
| 5426 |
return
|
| 5427 |
( a.low == b.low ) |
| 5428 |
&& ( ( a.high == b.high ) |
| 5429 |
|| ( ( a.low == 0 )
|
| 5430 |
&& ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) |
| 5431 |
); |
| 5432 |
|
| 5433 |
} |
| 5434 |
|
| 5435 |
/*----------------------------------------------------------------------------
|
| 5436 |
| Returns 1 if the quadruple-precision floating-point value `a' is less than
|
| 5437 |
| or equal to the corresponding value `b', and 0 otherwise. Quiet NaNs do not
|
| 5438 |
| cause an exception. Otherwise, the comparison is performed according to the
|
| 5439 |
| IEC/IEEE Standard for Binary Floating-Point Arithmetic.
|
| 5440 |
*----------------------------------------------------------------------------*/
|
| 5441 |
|
| 5442 |
int float128_le_quiet( float128 a, float128 b STATUS_PARAM )
|
| 5443 |
{
|
| 5444 |
flag aSign, bSign; |
| 5445 |
|
| 5446 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5447 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5448 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5449 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5450 |
) {
|
| 5451 |
if ( float128_is_signaling_nan( a )
|
| 5452 |
|| float128_is_signaling_nan( b ) ) {
|
| 5453 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5454 |
} |
| 5455 |
return 0; |
| 5456 |
} |
| 5457 |
aSign = extractFloat128Sign( a ); |
| 5458 |
bSign = extractFloat128Sign( b ); |
| 5459 |
if ( aSign != bSign ) {
|
| 5460 |
return
|
| 5461 |
aSign |
| 5462 |
|| ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 5463 |
== 0 );
|
| 5464 |
} |
| 5465 |
return
|
| 5466 |
aSign ? le128( b.high, b.low, a.high, a.low ) |
| 5467 |
: le128( a.high, a.low, b.high, b.low ); |
| 5468 |
|
| 5469 |
} |
| 5470 |
|
| 5471 |
/*----------------------------------------------------------------------------
|
| 5472 |
| Returns 1 if the quadruple-precision floating-point value `a' is less than
|
| 5473 |
| the corresponding value `b', and 0 otherwise. Quiet NaNs do not cause an
|
| 5474 |
| exception. Otherwise, the comparison is performed according to the IEC/IEEE
|
| 5475 |
| Standard for Binary Floating-Point Arithmetic.
|
| 5476 |
*----------------------------------------------------------------------------*/
|
| 5477 |
|
| 5478 |
int float128_lt_quiet( float128 a, float128 b STATUS_PARAM )
|
| 5479 |
{
|
| 5480 |
flag aSign, bSign; |
| 5481 |
|
| 5482 |
if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) |
| 5483 |
&& ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) |
| 5484 |
|| ( ( extractFloat128Exp( b ) == 0x7FFF )
|
| 5485 |
&& ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) |
| 5486 |
) {
|
| 5487 |
if ( float128_is_signaling_nan( a )
|
| 5488 |
|| float128_is_signaling_nan( b ) ) {
|
| 5489 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5490 |
} |
| 5491 |
return 0; |
| 5492 |
} |
| 5493 |
aSign = extractFloat128Sign( a ); |
| 5494 |
bSign = extractFloat128Sign( b ); |
| 5495 |
if ( aSign != bSign ) {
|
| 5496 |
return
|
| 5497 |
aSign |
| 5498 |
&& ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low )
|
| 5499 |
!= 0 );
|
| 5500 |
} |
| 5501 |
return
|
| 5502 |
aSign ? lt128( b.high, b.low, a.high, a.low ) |
| 5503 |
: lt128( a.high, a.low, b.high, b.low ); |
| 5504 |
|
| 5505 |
} |
| 5506 |
|
| 5507 |
#endif
|
| 5508 |
|
| 5509 |
/* misc functions */
|
| 5510 |
float32 uint32_to_float32( unsigned int a STATUS_PARAM ) |
| 5511 |
{
|
| 5512 |
return int64_to_float32(a STATUS_VAR);
|
| 5513 |
} |
| 5514 |
|
| 5515 |
float64 uint32_to_float64( unsigned int a STATUS_PARAM ) |
| 5516 |
{
|
| 5517 |
return int64_to_float64(a STATUS_VAR);
|
| 5518 |
} |
| 5519 |
|
| 5520 |
unsigned int float32_to_uint32( float32 a STATUS_PARAM ) |
| 5521 |
{
|
| 5522 |
int64_t v; |
| 5523 |
unsigned int res; |
| 5524 |
|
| 5525 |
v = float32_to_int64(a STATUS_VAR); |
| 5526 |
if (v < 0) { |
| 5527 |
res = 0;
|
| 5528 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5529 |
} else if (v > 0xffffffff) { |
| 5530 |
res = 0xffffffff;
|
| 5531 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5532 |
} else {
|
| 5533 |
res = v; |
| 5534 |
} |
| 5535 |
return res;
|
| 5536 |
} |
| 5537 |
|
| 5538 |
unsigned int float32_to_uint32_round_to_zero( float32 a STATUS_PARAM ) |
| 5539 |
{
|
| 5540 |
int64_t v; |
| 5541 |
unsigned int res; |
| 5542 |
|
| 5543 |
v = float32_to_int64_round_to_zero(a STATUS_VAR); |
| 5544 |
if (v < 0) { |
| 5545 |
res = 0;
|
| 5546 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5547 |
} else if (v > 0xffffffff) { |
| 5548 |
res = 0xffffffff;
|
| 5549 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5550 |
} else {
|
| 5551 |
res = v; |
| 5552 |
} |
| 5553 |
return res;
|
| 5554 |
} |
| 5555 |
|
| 5556 |
unsigned int float64_to_uint32( float64 a STATUS_PARAM ) |
| 5557 |
{
|
| 5558 |
int64_t v; |
| 5559 |
unsigned int res; |
| 5560 |
|
| 5561 |
v = float64_to_int64(a STATUS_VAR); |
| 5562 |
if (v < 0) { |
| 5563 |
res = 0;
|
| 5564 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5565 |
} else if (v > 0xffffffff) { |
| 5566 |
res = 0xffffffff;
|
| 5567 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5568 |
} else {
|
| 5569 |
res = v; |
| 5570 |
} |
| 5571 |
return res;
|
| 5572 |
} |
| 5573 |
|
| 5574 |
unsigned int float64_to_uint32_round_to_zero( float64 a STATUS_PARAM ) |
| 5575 |
{
|
| 5576 |
int64_t v; |
| 5577 |
unsigned int res; |
| 5578 |
|
| 5579 |
v = float64_to_int64_round_to_zero(a STATUS_VAR); |
| 5580 |
if (v < 0) { |
| 5581 |
res = 0;
|
| 5582 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5583 |
} else if (v > 0xffffffff) { |
| 5584 |
res = 0xffffffff;
|
| 5585 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5586 |
} else {
|
| 5587 |
res = v; |
| 5588 |
} |
| 5589 |
return res;
|
| 5590 |
} |
| 5591 |
|
| 5592 |
/* FIXME: This looks broken. */
|
| 5593 |
uint64_t float64_to_uint64 (float64 a STATUS_PARAM) |
| 5594 |
{
|
| 5595 |
int64_t v; |
| 5596 |
|
| 5597 |
v = float64_val(int64_to_float64(INT64_MIN STATUS_VAR)); |
| 5598 |
v += float64_val(a); |
| 5599 |
v = float64_to_int64(make_float64(v) STATUS_VAR); |
| 5600 |
|
| 5601 |
return v - INT64_MIN;
|
| 5602 |
} |
| 5603 |
|
| 5604 |
uint64_t float64_to_uint64_round_to_zero (float64 a STATUS_PARAM) |
| 5605 |
{
|
| 5606 |
int64_t v; |
| 5607 |
|
| 5608 |
v = float64_val(int64_to_float64(INT64_MIN STATUS_VAR)); |
| 5609 |
v += float64_val(a); |
| 5610 |
v = float64_to_int64_round_to_zero(make_float64(v) STATUS_VAR); |
| 5611 |
|
| 5612 |
return v - INT64_MIN;
|
| 5613 |
} |
| 5614 |
|
| 5615 |
#define COMPARE(s, nan_exp) \
|
| 5616 |
INLINE int float ## s ## _compare_internal( float ## s a, float ## s b, \ |
| 5617 |
int is_quiet STATUS_PARAM ) \
|
| 5618 |
{ \
|
| 5619 |
flag aSign, bSign; \ |
| 5620 |
bits ## s av, bv; \ |
| 5621 |
\ |
| 5622 |
if (( ( extractFloat ## s ## Exp( a ) == nan_exp ) && \ |
| 5623 |
extractFloat ## s ## Frac( a ) ) || \ |
| 5624 |
( ( extractFloat ## s ## Exp( b ) == nan_exp ) && \ |
| 5625 |
extractFloat ## s ## Frac( b ) )) { \ |
| 5626 |
if (!is_quiet || \
|
| 5627 |
float ## s ## _is_signaling_nan( a ) || \ |
| 5628 |
float ## s ## _is_signaling_nan( b ) ) { \ |
| 5629 |
float_raise( float_flag_invalid STATUS_VAR); \ |
| 5630 |
} \ |
| 5631 |
return float_relation_unordered; \
|
| 5632 |
} \ |
| 5633 |
aSign = extractFloat ## s ## Sign( a ); \ |
| 5634 |
bSign = extractFloat ## s ## Sign( b ); \ |
| 5635 |
av = float ## s ## _val(a); \ |
| 5636 |
bv = float ## s ## _val(b); \ |
| 5637 |
if ( aSign != bSign ) { \
|
| 5638 |
if ( (bits ## s) ( ( av | bv )<<1 ) == 0 ) { \ |
| 5639 |
/* zero case */ \
|
| 5640 |
return float_relation_equal; \
|
| 5641 |
} else { \
|
| 5642 |
return 1 - (2 * aSign); \ |
| 5643 |
} \ |
| 5644 |
} else { \
|
| 5645 |
if (av == bv) { \
|
| 5646 |
return float_relation_equal; \
|
| 5647 |
} else { \
|
| 5648 |
return 1 - 2 * (aSign ^ ( av < bv )); \ |
| 5649 |
} \ |
| 5650 |
} \ |
| 5651 |
} \ |
| 5652 |
\ |
| 5653 |
int float ## s ## _compare( float ## s a, float ## s b STATUS_PARAM ) \ |
| 5654 |
{ \
|
| 5655 |
return float ## s ## _compare_internal(a, b, 0 STATUS_VAR); \ |
| 5656 |
} \ |
| 5657 |
\ |
| 5658 |
int float ## s ## _compare_quiet( float ## s a, float ## s b STATUS_PARAM ) \ |
| 5659 |
{ \
|
| 5660 |
return float ## s ## _compare_internal(a, b, 1 STATUS_VAR); \ |
| 5661 |
} |
| 5662 |
|
| 5663 |
COMPARE(32, 0xff) |
| 5664 |
COMPARE(64, 0x7ff) |
| 5665 |
|
| 5666 |
INLINE int float128_compare_internal( float128 a, float128 b,
|
| 5667 |
int is_quiet STATUS_PARAM )
|
| 5668 |
{
|
| 5669 |
flag aSign, bSign; |
| 5670 |
|
| 5671 |
if (( ( extractFloat128Exp( a ) == 0x7fff ) && |
| 5672 |
( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) || |
| 5673 |
( ( extractFloat128Exp( b ) == 0x7fff ) &&
|
| 5674 |
( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) )) {
|
| 5675 |
if (!is_quiet ||
|
| 5676 |
float128_is_signaling_nan( a ) || |
| 5677 |
float128_is_signaling_nan( b ) ) {
|
| 5678 |
float_raise( float_flag_invalid STATUS_VAR); |
| 5679 |
} |
| 5680 |
return float_relation_unordered;
|
| 5681 |
} |
| 5682 |
aSign = extractFloat128Sign( a ); |
| 5683 |
bSign = extractFloat128Sign( b ); |
| 5684 |
if ( aSign != bSign ) {
|
| 5685 |
if ( ( ( ( a.high | b.high )<<1 ) | a.low | b.low ) == 0 ) { |
| 5686 |
/* zero case */
|
| 5687 |
return float_relation_equal;
|
| 5688 |
} else {
|
| 5689 |
return 1 - (2 * aSign); |
| 5690 |
} |
| 5691 |
} else {
|
| 5692 |
if (a.low == b.low && a.high == b.high) {
|
| 5693 |
return float_relation_equal;
|
| 5694 |
} else {
|
| 5695 |
return 1 - 2 * (aSign ^ ( lt128( a.high, a.low, b.high, b.low ) )); |
| 5696 |
} |
| 5697 |
} |
| 5698 |
} |
| 5699 |
|
| 5700 |
int float128_compare( float128 a, float128 b STATUS_PARAM )
|
| 5701 |
{
|
| 5702 |
return float128_compare_internal(a, b, 0 STATUS_VAR); |
| 5703 |
} |
| 5704 |
|
| 5705 |
int float128_compare_quiet( float128 a, float128 b STATUS_PARAM )
|
| 5706 |
{
|
| 5707 |
return float128_compare_internal(a, b, 1 STATUS_VAR); |
| 5708 |
} |
| 5709 |
|
| 5710 |
/* Multiply A by 2 raised to the power N. */
|
| 5711 |
float32 float32_scalbn( float32 a, int n STATUS_PARAM )
|
| 5712 |
{
|
| 5713 |
flag aSign; |
| 5714 |
int16 aExp; |
| 5715 |
bits32 aSig; |
| 5716 |
|
| 5717 |
aSig = extractFloat32Frac( a ); |
| 5718 |
aExp = extractFloat32Exp( a ); |
| 5719 |
aSign = extractFloat32Sign( a ); |
| 5720 |
|
| 5721 |
if ( aExp == 0xFF ) { |
| 5722 |
return a;
|
| 5723 |
} |
| 5724 |
if ( aExp != 0 ) |
| 5725 |
aSig |= 0x00800000;
|
| 5726 |
else if ( aSig == 0 ) |
| 5727 |
return a;
|
| 5728 |
|
| 5729 |
aExp += n - 1;
|
| 5730 |
aSig <<= 7;
|
| 5731 |
return normalizeRoundAndPackFloat32( aSign, aExp, aSig STATUS_VAR );
|
| 5732 |
} |
| 5733 |
|
| 5734 |
float64 float64_scalbn( float64 a, int n STATUS_PARAM )
|
| 5735 |
{
|
| 5736 |
flag aSign; |
| 5737 |
int16 aExp; |
| 5738 |
bits64 aSig; |
| 5739 |
|
| 5740 |
aSig = extractFloat64Frac( a ); |
| 5741 |
aExp = extractFloat64Exp( a ); |
| 5742 |
aSign = extractFloat64Sign( a ); |
| 5743 |
|
| 5744 |
if ( aExp == 0x7FF ) { |
| 5745 |
return a;
|
| 5746 |
} |
| 5747 |
if ( aExp != 0 ) |
| 5748 |
aSig |= LIT64( 0x0010000000000000 );
|
| 5749 |
else if ( aSig == 0 ) |
| 5750 |
return a;
|
| 5751 |
|
| 5752 |
aExp += n - 1;
|
| 5753 |
aSig <<= 10;
|
| 5754 |
return normalizeRoundAndPackFloat64( aSign, aExp, aSig STATUS_VAR );
|
| 5755 |
} |
| 5756 |
|
| 5757 |
#ifdef FLOATX80
|
| 5758 |
floatx80 floatx80_scalbn( floatx80 a, int n STATUS_PARAM )
|
| 5759 |
{
|
| 5760 |
flag aSign; |
| 5761 |
int16 aExp; |
| 5762 |
bits64 aSig; |
| 5763 |
|
| 5764 |
aSig = extractFloatx80Frac( a ); |
| 5765 |
aExp = extractFloatx80Exp( a ); |
| 5766 |
aSign = extractFloatx80Sign( a ); |
| 5767 |
|
| 5768 |
if ( aExp == 0x7FF ) { |
| 5769 |
return a;
|
| 5770 |
} |
| 5771 |
if (aExp == 0 && aSig == 0) |
| 5772 |
return a;
|
| 5773 |
|
| 5774 |
aExp += n; |
| 5775 |
return normalizeRoundAndPackFloatx80( STATUS(floatx80_rounding_precision),
|
| 5776 |
aSign, aExp, aSig, 0 STATUS_VAR );
|
| 5777 |
} |
| 5778 |
#endif
|
| 5779 |
|
| 5780 |
#ifdef FLOAT128
|
| 5781 |
float128 float128_scalbn( float128 a, int n STATUS_PARAM )
|
| 5782 |
{
|
| 5783 |
flag aSign; |
| 5784 |
int32 aExp; |
| 5785 |
bits64 aSig0, aSig1; |
| 5786 |
|
| 5787 |
aSig1 = extractFloat128Frac1( a ); |
| 5788 |
aSig0 = extractFloat128Frac0( a ); |
| 5789 |
aExp = extractFloat128Exp( a ); |
| 5790 |
aSign = extractFloat128Sign( a ); |
| 5791 |
if ( aExp == 0x7FFF ) { |
| 5792 |
return a;
|
| 5793 |
} |
| 5794 |
if ( aExp != 0 ) |
| 5795 |
aSig0 |= LIT64( 0x0001000000000000 );
|
| 5796 |
else if ( aSig0 == 0 && aSig1 == 0 ) |
| 5797 |
return a;
|
| 5798 |
|
| 5799 |
aExp += n - 1;
|
| 5800 |
return normalizeRoundAndPackFloat128( aSign, aExp, aSig0, aSig1
|
| 5801 |
STATUS_VAR ); |
| 5802 |
|
| 5803 |
} |
| 5804 |
#endif
|