diff options
Diffstat (limited to 'src/lib/softfloat')
| -rwxr-xr-x | src/lib/softfloat/fsincos.c | 611 | ||||
| -rw-r--r-- | src/lib/softfloat/fyl2x.c | 468 | ||||
| -rw-r--r-- | src/lib/softfloat/softfloat.c | 6589 | ||||
| -rw-r--r-- | src/lib/softfloat/softfloat.h | 259 |
4 files changed, 3962 insertions, 3965 deletions
diff --git a/src/lib/softfloat/fsincos.c b/src/lib/softfloat/fsincos.c index fb148cc8fbe..84c67c47270 100755 --- a/src/lib/softfloat/fsincos.c +++ b/src/lib/softfloat/fsincos.c @@ -47,7 +47,7 @@ static const floatx80 floatx80_default_nan = packFloatx80(0, 0xffff, U64(0xfffff INLINE bits64 extractFloatx80Frac( floatx80 a ) { - return a.low; + return a.low; } @@ -59,7 +59,7 @@ INLINE bits64 extractFloatx80Frac( floatx80 a ) INLINE int32 extractFloatx80Exp( floatx80 a ) { - return a.high & 0x7FFF; + return a.high & 0x7FFF; } @@ -71,7 +71,7 @@ INLINE int32 extractFloatx80Exp( floatx80 a ) INLINE flag extractFloatx80Sign( floatx80 a ) { - return a.high>>15; + return a.high>>15; } @@ -83,12 +83,12 @@ INLINE flag extractFloatx80Sign( floatx80 a ) INLINE floatx80 propagateFloatx80NaNOneArg(floatx80 a) { - if (floatx80_is_signaling_nan(a)) - float_raise(float_flag_invalid); + if (floatx80_is_signaling_nan(a)) + float_raise(float_flag_invalid); - a.low |= U64(0xC000000000000000); + a.low |= U64(0xC000000000000000); - return a; + return a; } /*---------------------------------------------------------------------------- @@ -100,62 +100,62 @@ INLINE floatx80 propagateFloatx80NaNOneArg(floatx80 a) void normalizeFloatx80Subnormal(UINT64 aSig, INT32 *zExpPtr, UINT64 *zSigPtr) { - int shiftCount = countLeadingZeros64(aSig); - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; + int shiftCount = countLeadingZeros64(aSig); + *zSigPtr = aSig<<shiftCount; + *zExpPtr = 1 - shiftCount; } /* reduce trigonometric function argument using 128-bit precision M_PI approximation */ static UINT64 argument_reduction_kernel(UINT64 aSig0, int Exp, UINT64 *zSig0, UINT64 *zSig1) { - UINT64 term0, term1, term2; - UINT64 aSig1 = 0; - - shortShift128Left(aSig1, aSig0, Exp, &aSig1, &aSig0); - UINT64 q = estimateDiv128To64(aSig1, aSig0, FLOAT_PI_HI); - mul128By64To192(FLOAT_PI_HI, FLOAT_PI_LO, q, &term0, &term1, &term2); - sub128(aSig1, aSig0, term0, term1, zSig1, zSig0); - while ((INT64)(*zSig1) < 0) { - --q; - add192(*zSig1, *zSig0, term2, 0, FLOAT_PI_HI, FLOAT_PI_LO, zSig1, zSig0, &term2); - } - *zSig1 = term2; - return q; + UINT64 term0, term1, term2; + UINT64 aSig1 = 0; + + shortShift128Left(aSig1, aSig0, Exp, &aSig1, &aSig0); + UINT64 q = estimateDiv128To64(aSig1, aSig0, FLOAT_PI_HI); + mul128By64To192(FLOAT_PI_HI, FLOAT_PI_LO, q, &term0, &term1, &term2); + sub128(aSig1, aSig0, term0, term1, zSig1, zSig0); + while ((INT64)(*zSig1) < 0) { + --q; + add192(*zSig1, *zSig0, term2, 0, FLOAT_PI_HI, FLOAT_PI_LO, zSig1, zSig0, &term2); + } + *zSig1 = term2; + return q; } static int reduce_trig_arg(int expDiff, int &zSign, UINT64 &aSig0, UINT64 &aSig1) { - UINT64 term0, term1, q = 0; - - if (expDiff < 0) { - shift128Right(aSig0, 0, 1, &aSig0, &aSig1); - expDiff = 0; - } - if (expDiff > 0) { - q = argument_reduction_kernel(aSig0, expDiff, &aSig0, &aSig1); - } - else { - if (FLOAT_PI_HI <= aSig0) { - aSig0 -= FLOAT_PI_HI; - q = 1; - } - } - - shift128Right(FLOAT_PI_HI, FLOAT_PI_LO, 1, &term0, &term1); - if (! lt128(aSig0, aSig1, term0, term1)) - { - int lt = lt128(term0, term1, aSig0, aSig1); - int eq = eq128(aSig0, aSig1, term0, term1); - - if ((eq && (q & 1)) || lt) { - zSign = !zSign; - ++q; - } - if (lt) sub128(FLOAT_PI_HI, FLOAT_PI_LO, aSig0, aSig1, &aSig0, &aSig1); - } - - return (int)(q & 3); + UINT64 term0, term1, q = 0; + + if (expDiff < 0) { + shift128Right(aSig0, 0, 1, &aSig0, &aSig1); + expDiff = 0; + } + if (expDiff > 0) { + q = argument_reduction_kernel(aSig0, expDiff, &aSig0, &aSig1); + } + else { + if (FLOAT_PI_HI <= aSig0) { + aSig0 -= FLOAT_PI_HI; + q = 1; + } + } + + shift128Right(FLOAT_PI_HI, FLOAT_PI_LO, 1, &term0, &term1); + if (! lt128(aSig0, aSig1, term0, term1)) + { + int lt = lt128(term0, term1, aSig0, aSig1); + int eq = eq128(aSig0, aSig1, term0, term1); + + if ((eq && (q & 1)) || lt) { + zSign = !zSign; + ++q; + } + if (lt) sub128(FLOAT_PI_HI, FLOAT_PI_LO, aSig0, aSig1, &aSig0, &aSig1); + } + + return (int)(q & 3); } #define SIN_ARR_SIZE 11 @@ -163,32 +163,32 @@ static int reduce_trig_arg(int expDiff, int &zSign, UINT64 &aSig0, UINT64 &aSig1 static float128 sin_arr[SIN_ARR_SIZE] = { - PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 1 */ - PACK_FLOAT_128(0xbffc555555555555, 0x5555555555555555), /* 3 */ - PACK_FLOAT_128(0x3ff8111111111111, 0x1111111111111111), /* 5 */ - PACK_FLOAT_128(0xbff2a01a01a01a01, 0xa01a01a01a01a01a), /* 7 */ - PACK_FLOAT_128(0x3fec71de3a556c73, 0x38faac1c88e50017), /* 9 */ - PACK_FLOAT_128(0xbfe5ae64567f544e, 0x38fe747e4b837dc7), /* 11 */ - PACK_FLOAT_128(0x3fde6124613a86d0, 0x97ca38331d23af68), /* 13 */ - PACK_FLOAT_128(0xbfd6ae7f3e733b81, 0xf11d8656b0ee8cb0), /* 15 */ - PACK_FLOAT_128(0x3fce952c77030ad4, 0xa6b2605197771b00), /* 17 */ - PACK_FLOAT_128(0xbfc62f49b4681415, 0x724ca1ec3b7b9675), /* 19 */ - PACK_FLOAT_128(0x3fbd71b8ef6dcf57, 0x18bef146fcee6e45) /* 21 */ + PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 1 */ + PACK_FLOAT_128(0xbffc555555555555, 0x5555555555555555), /* 3 */ + PACK_FLOAT_128(0x3ff8111111111111, 0x1111111111111111), /* 5 */ + PACK_FLOAT_128(0xbff2a01a01a01a01, 0xa01a01a01a01a01a), /* 7 */ + PACK_FLOAT_128(0x3fec71de3a556c73, 0x38faac1c88e50017), /* 9 */ + PACK_FLOAT_128(0xbfe5ae64567f544e, 0x38fe747e4b837dc7), /* 11 */ + PACK_FLOAT_128(0x3fde6124613a86d0, 0x97ca38331d23af68), /* 13 */ + PACK_FLOAT_128(0xbfd6ae7f3e733b81, 0xf11d8656b0ee8cb0), /* 15 */ + PACK_FLOAT_128(0x3fce952c77030ad4, 0xa6b2605197771b00), /* 17 */ + PACK_FLOAT_128(0xbfc62f49b4681415, 0x724ca1ec3b7b9675), /* 19 */ + PACK_FLOAT_128(0x3fbd71b8ef6dcf57, 0x18bef146fcee6e45) /* 21 */ }; static float128 cos_arr[COS_ARR_SIZE] = { - PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 0 */ - PACK_FLOAT_128(0xbffe000000000000, 0x0000000000000000), /* 2 */ - PACK_FLOAT_128(0x3ffa555555555555, 0x5555555555555555), /* 4 */ - PACK_FLOAT_128(0xbff56c16c16c16c1, 0x6c16c16c16c16c17), /* 6 */ - PACK_FLOAT_128(0x3fefa01a01a01a01, 0xa01a01a01a01a01a), /* 8 */ - PACK_FLOAT_128(0xbfe927e4fb7789f5, 0xc72ef016d3ea6679), /* 10 */ - PACK_FLOAT_128(0x3fe21eed8eff8d89, 0x7b544da987acfe85), /* 12 */ - PACK_FLOAT_128(0xbfda93974a8c07c9, 0xd20badf145dfa3e5), /* 14 */ - PACK_FLOAT_128(0x3fd2ae7f3e733b81, 0xf11d8656b0ee8cb0), /* 16 */ - PACK_FLOAT_128(0xbfca6827863b97d9, 0x77bb004886a2c2ab), /* 18 */ - PACK_FLOAT_128(0x3fc1e542ba402022, 0x507a9cad2bf8f0bb) /* 20 */ + PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 0 */ + PACK_FLOAT_128(0xbffe000000000000, 0x0000000000000000), /* 2 */ + PACK_FLOAT_128(0x3ffa555555555555, 0x5555555555555555), /* 4 */ + PACK_FLOAT_128(0xbff56c16c16c16c1, 0x6c16c16c16c16c17), /* 6 */ + PACK_FLOAT_128(0x3fefa01a01a01a01, 0xa01a01a01a01a01a), /* 8 */ + PACK_FLOAT_128(0xbfe927e4fb7789f5, 0xc72ef016d3ea6679), /* 10 */ + PACK_FLOAT_128(0x3fe21eed8eff8d89, 0x7b544da987acfe85), /* 12 */ + PACK_FLOAT_128(0xbfda93974a8c07c9, 0xd20badf145dfa3e5), /* 14 */ + PACK_FLOAT_128(0x3fd2ae7f3e733b81, 0xf11d8656b0ee8cb0), /* 16 */ + PACK_FLOAT_128(0xbfca6827863b97d9, 0x77bb004886a2c2ab), /* 18 */ + PACK_FLOAT_128(0x3fc1e542ba402022, 0x507a9cad2bf8f0bb) /* 20 */ }; extern float128 OddPoly (float128 x, float128 *arr, unsigned n); @@ -196,27 +196,27 @@ extern float128 OddPoly (float128 x, float128 *arr, unsigned n); /* 0 <= x <= pi/4 */ INLINE float128 poly_sin(float128 x) { - // 3 5 7 9 11 13 15 - // x x x x x x x - // sin (x) ~ x - --- + --- - --- + --- - ---- + ---- - ---- = - // 3! 5! 7! 9! 11! 13! 15! - // - // 2 4 6 8 10 12 14 - // x x x x x x x - // = x * [ 1 - --- + --- - --- + --- - ---- + ---- - ---- ] = - // 3! 5! 7! 9! 11! 13! 15! - // - // 3 3 - // -- 4k -- 4k+2 - // p(x) = > C * x > 0 q(x) = > C * x < 0 - // -- 2k -- 2k+1 - // k=0 k=0 - // - // 2 - // sin(x) ~ x * [ p(x) + x * q(x) ] - // - - return OddPoly(x, sin_arr, SIN_ARR_SIZE); + // 3 5 7 9 11 13 15 + // x x x x x x x + // sin (x) ~ x - --- + --- - --- + --- - ---- + ---- - ---- = + // 3! 5! 7! 9! 11! 13! 15! + // + // 2 4 6 8 10 12 14 + // x x x x x x x + // = x * [ 1 - --- + --- - --- + --- - ---- + ---- - ---- ] = + // 3! 5! 7! 9! 11! 13! 15! + // + // 3 3 + // -- 4k -- 4k+2 + // p(x) = > C * x > 0 q(x) = > C * x < 0 + // -- 2k -- 2k+1 + // k=0 k=0 + // + // 2 + // sin(x) ~ x * [ p(x) + x * q(x) ] + // + + return OddPoly(x, sin_arr, SIN_ARR_SIZE); } extern float128 EvenPoly(float128 x, float128 *arr, unsigned n); @@ -224,53 +224,53 @@ extern float128 EvenPoly(float128 x, float128 *arr, unsigned n); /* 0 <= x <= pi/4 */ INLINE float128 poly_cos(float128 x) { - // 2 4 6 8 10 12 14 - // x x x x x x x - // cos (x) ~ 1 - --- + --- - --- + --- - ---- + ---- - ---- - // 2! 4! 6! 8! 10! 12! 14! - // - // 3 3 - // -- 4k -- 4k+2 - // p(x) = > C * x > 0 q(x) = > C * x < 0 - // -- 2k -- 2k+1 - // k=0 k=0 - // - // 2 - // cos(x) ~ [ p(x) + x * q(x) ] - // - - return EvenPoly(x, cos_arr, COS_ARR_SIZE); + // 2 4 6 8 10 12 14 + // x x x x x x x + // cos (x) ~ 1 - --- + --- - --- + --- - ---- + ---- - ---- + // 2! 4! 6! 8! 10! 12! 14! + // + // 3 3 + // -- 4k -- 4k+2 + // p(x) = > C * x > 0 q(x) = > C * x < 0 + // -- 2k -- 2k+1 + // k=0 k=0 + // + // 2 + // cos(x) ~ [ p(x) + x * q(x) ] + // + + return EvenPoly(x, cos_arr, COS_ARR_SIZE); } INLINE void sincos_invalid(floatx80 *sin_a, floatx80 *cos_a, floatx80 a) { - if (sin_a) *sin_a = a; - if (cos_a) *cos_a = a; + if (sin_a) *sin_a = a; + if (cos_a) *cos_a = a; } INLINE void sincos_tiny_argument(floatx80 *sin_a, floatx80 *cos_a, floatx80 a) { - if (sin_a) *sin_a = a; - if (cos_a) *cos_a = floatx80_one; + if (sin_a) *sin_a = a; + if (cos_a) *cos_a = floatx80_one; } static floatx80 sincos_approximation(int neg, float128 r, UINT64 quotient) { - if (quotient & 0x1) { - r = poly_cos(r); - neg = 0; - } else { - r = poly_sin(r); - } + if (quotient & 0x1) { + r = poly_cos(r); + neg = 0; + } else { + r = poly_sin(r); + } - floatx80 result = float128_to_floatx80(r); - if (quotient & 0x2) - neg = ! neg; + floatx80 result = float128_to_floatx80(r); + if (quotient & 0x2) + neg = ! neg; - if (neg) - result = floatx80_chs(result); + if (neg) + result = floatx80_chs(result); - return result; + return result; } // ================================================= @@ -295,92 +295,92 @@ static floatx80 sincos_approximation(int neg, float128 r, UINT64 quotient) int sf_fsincos(floatx80 a, floatx80 *sin_a, floatx80 *cos_a) { - UINT64 aSig0, aSig1 = 0; - INT32 aExp, zExp, expDiff; - int aSign, zSign; - int q = 0; - - aSig0 = extractFloatx80Frac(a); - aExp = extractFloatx80Exp(a); - aSign = extractFloatx80Sign(a); - - /* invalid argument */ - if (aExp == 0x7FFF) { - if ((UINT64) (aSig0<<1)) { - sincos_invalid(sin_a, cos_a, propagateFloatx80NaNOneArg(a)); - return 0; - } - - float_raise(float_flag_invalid); - sincos_invalid(sin_a, cos_a, floatx80_default_nan); - return 0; - } - - if (aExp == 0) { - if (aSig0 == 0) { - sincos_tiny_argument(sin_a, cos_a, a); - return 0; - } + UINT64 aSig0, aSig1 = 0; + INT32 aExp, zExp, expDiff; + int aSign, zSign; + int q = 0; + + aSig0 = extractFloatx80Frac(a); + aExp = extractFloatx80Exp(a); + aSign = extractFloatx80Sign(a); + + /* invalid argument */ + if (aExp == 0x7FFF) { + if ((UINT64) (aSig0<<1)) { + sincos_invalid(sin_a, cos_a, propagateFloatx80NaNOneArg(a)); + return 0; + } + + float_raise(float_flag_invalid); + sincos_invalid(sin_a, cos_a, floatx80_default_nan); + return 0; + } + + if (aExp == 0) { + if (aSig0 == 0) { + sincos_tiny_argument(sin_a, cos_a, a); + return 0; + } // float_raise(float_flag_denormal); - /* handle pseudo denormals */ - if (! (aSig0 & U64(0x8000000000000000))) - { - float_raise(float_flag_inexact); - if (sin_a) - float_raise(float_flag_underflow); - sincos_tiny_argument(sin_a, cos_a, a); - return 0; - } - - normalizeFloatx80Subnormal(aSig0, &aExp, &aSig0); - } - - zSign = aSign; - zExp = EXP_BIAS; - expDiff = aExp - zExp; - - /* argument is out-of-range */ - if (expDiff >= 63) - return -1; - - float_raise(float_flag_inexact); - - if (expDiff < -1) { // doesn't require reduction - if (expDiff <= -68) { - a = packFloatx80(aSign, aExp, aSig0); - sincos_tiny_argument(sin_a, cos_a, a); - return 0; - } - zExp = aExp; - } - else { - q = reduce_trig_arg(expDiff, zSign, aSig0, aSig1); - } - - /* **************************** */ - /* argument reduction completed */ - /* **************************** */ - - /* using float128 for approximation */ - float128 r = normalizeRoundAndPackFloat128(0, zExp-0x10, aSig0, aSig1); - - if (aSign) q = -q; - if (sin_a) *sin_a = sincos_approximation(zSign, r, q); - if (cos_a) *cos_a = sincos_approximation(zSign, r, q+1); - - return 0; + /* handle pseudo denormals */ + if (! (aSig0 & U64(0x8000000000000000))) + { + float_raise(float_flag_inexact); + if (sin_a) + float_raise(float_flag_underflow); + sincos_tiny_argument(sin_a, cos_a, a); + return 0; + } + + normalizeFloatx80Subnormal(aSig0, &aExp, &aSig0); + } + + zSign = aSign; + zExp = EXP_BIAS; + expDiff = aExp - zExp; + + /* argument is out-of-range */ + if (expDiff >= 63) + return -1; + + float_raise(float_flag_inexact); + + if (expDiff < -1) { // doesn't require reduction + if (expDiff <= -68) { + a = packFloatx80(aSign, aExp, aSig0); + sincos_tiny_argument(sin_a, cos_a, a); + return 0; + } + zExp = aExp; + } + else { + q = reduce_trig_arg(expDiff, zSign, aSig0, aSig1); + } + + /* **************************** */ + /* argument reduction completed */ + /* **************************** */ + + /* using float128 for approximation */ + float128 r = normalizeRoundAndPackFloat128(0, zExp-0x10, aSig0, aSig1); + + if (aSign) q = -q; + if (sin_a) *sin_a = sincos_approximation(zSign, r, q); + if (cos_a) *cos_a = sincos_approximation(zSign, r, q+1); + + return 0; } int floatx80_fsin(floatx80 &a) { - return sf_fsincos(a, &a, 0); + return sf_fsincos(a, &a, 0); } int floatx80_fcos(floatx80 &a) { - return sf_fsincos(a, 0, &a); + return sf_fsincos(a, 0, &a); } // ================================================= @@ -412,83 +412,83 @@ int floatx80_fcos(floatx80 &a) int floatx80_ftan(floatx80 &a) { - UINT64 aSig0, aSig1 = 0; - INT32 aExp, zExp, expDiff; - int aSign, zSign; - int q = 0; - - aSig0 = extractFloatx80Frac(a); - aExp = extractFloatx80Exp(a); - aSign = extractFloatx80Sign(a); - - /* invalid argument */ - if (aExp == 0x7FFF) { - if ((UINT64) (aSig0<<1)) - { - a = propagateFloatx80NaNOneArg(a); - return 0; - } - - float_raise(float_flag_invalid); - a = floatx80_default_nan; - return 0; - } - - if (aExp == 0) { - if (aSig0 == 0) return 0; + UINT64 aSig0, aSig1 = 0; + INT32 aExp, zExp, expDiff; + int aSign, zSign; + int q = 0; + + aSig0 = extractFloatx80Frac(a); + aExp = extractFloatx80Exp(a); + aSign = extractFloatx80Sign(a); + + /* invalid argument */ + if (aExp == 0x7FFF) { + if ((UINT64) (aSig0<<1)) + { + a = propagateFloatx80NaNOneArg(a); + return 0; + } + + float_raise(float_flag_invalid); + a = floatx80_default_nan; + return 0; + } + + if (aExp == 0) { + if (aSig0 == 0) return 0; // float_raise(float_flag_denormal); - /* handle pseudo denormals */ - if (! (aSig0 & U64(0x8000000000000000))) - { - float_raise(float_flag_inexact | float_flag_underflow); - return 0; - } - normalizeFloatx80Subnormal(aSig0, &aExp, &aSig0); - } - - zSign = aSign; - zExp = EXP_BIAS; - expDiff = aExp - zExp; - - /* argument is out-of-range */ - if (expDiff >= 63) - return -1; - - float_raise(float_flag_inexact); - - if (expDiff < -1) { // doesn't require reduction - if (expDiff <= -68) { - a = packFloatx80(aSign, aExp, aSig0); - return 0; - } - zExp = aExp; - } - else { - q = reduce_trig_arg(expDiff, zSign, aSig0, aSig1); - } - - /* **************************** */ - /* argument reduction completed */ - /* **************************** */ - - /* using float128 for approximation */ - float128 r = normalizeRoundAndPackFloat128(0, zExp-0x10, aSig0, aSig1); - - float128 sin_r = poly_sin(r); - float128 cos_r = poly_cos(r); - - if (q & 0x1) { - r = float128_div(cos_r, sin_r); - zSign = ! zSign; - } else { - r = float128_div(sin_r, cos_r); - } - - a = float128_to_floatx80(r); - if (zSign) - a = floatx80_chs(a); - - return 0; + /* handle pseudo denormals */ + if (! (aSig0 & U64(0x8000000000000000))) + { + float_raise(float_flag_inexact | float_flag_underflow); + return 0; + } + normalizeFloatx80Subnormal(aSig0, &aExp, &aSig0); + } + + zSign = aSign; + zExp = EXP_BIAS; + expDiff = aExp - zExp; + + /* argument is out-of-range */ + if (expDiff >= 63) + return -1; + + float_raise(float_flag_inexact); + + if (expDiff < -1) { // doesn't require reduction + if (expDiff <= -68) { + a = packFloatx80(aSign, aExp, aSig0); + return 0; + } + zExp = aExp; + } + else { + q = reduce_trig_arg(expDiff, zSign, aSig0, aSig1); + } + + /* **************************** */ + /* argument reduction completed */ + /* **************************** */ + + /* using float128 for approximation */ + float128 r = normalizeRoundAndPackFloat128(0, zExp-0x10, aSig0, aSig1); + + float128 sin_r = poly_sin(r); + float128 cos_r = poly_cos(r); + + if (q & 0x1) { + r = float128_div(cos_r, sin_r); + zSign = ! zSign; + } else { + r = float128_div(sin_r, cos_r); + } + + a = float128_to_floatx80(r); + if (zSign) + a = floatx80_chs(a); + + return 0; } // 2 3 4 n @@ -504,30 +504,30 @@ int floatx80_ftan(floatx80 &a) float128 EvalPoly(float128 x, float128 *arr, unsigned n) { - float128 x2 = float128_mul(x, x); - unsigned i; - - assert(n > 1); - - float128 r1 = arr[--n]; - i = n; - while(i >= 2) { - r1 = float128_mul(r1, x2); - i -= 2; - r1 = float128_add(r1, arr[i]); - } - if (i) r1 = float128_mul(r1, x); - - float128 r2 = arr[--n]; - i = n; - while(i >= 2) { - r2 = float128_mul(r2, x2); - i -= 2; - r2 = float128_add(r2, arr[i]); - } - if (i) r2 = float128_mul(r2, x); - - return float128_add(r1, r2); + float128 x2 = float128_mul(x, x); + unsigned i; + + assert(n > 1); + + float128 r1 = arr[--n]; + i = n; + while(i >= 2) { + r1 = float128_mul(r1, x2); + i -= 2; + r1 = float128_add(r1, arr[i]); + } + if (i) r1 = float128_mul(r1, x); + + float128 r2 = arr[--n]; + i = n; + while(i >= 2) { + r2 = float128_mul(r2, x2); + i -= 2; + r2 = float128_add(r2, arr[i]); + } + if (i) r2 = float128_mul(r2, x); + + return float128_add(r1, r2); } // 2 4 6 8 2n @@ -544,7 +544,7 @@ float128 EvalPoly(float128 x, float128 *arr, unsigned n) float128 EvenPoly(float128 x, float128 *arr, unsigned n) { - return EvalPoly(float128_mul(x, x), arr, n); + return EvalPoly(float128_mul(x, x), arr, n); } // 3 5 7 9 2n+1 @@ -564,6 +564,5 @@ float128 EvenPoly(float128 x, float128 *arr, unsigned n) float128 OddPoly(float128 x, float128 *arr, unsigned n) { - return float128_mul(x, EvenPoly(x, arr, n)); + return float128_mul(x, EvenPoly(x, arr, n)); } - diff --git a/src/lib/softfloat/fyl2x.c b/src/lib/softfloat/fyl2x.c index fc41d125581..be5e134cc3b 100644 --- a/src/lib/softfloat/fyl2x.c +++ b/src/lib/softfloat/fyl2x.c @@ -53,7 +53,7 @@ static const floatx80 floatx80_default_nan = packFloatx80(0, 0xffff, U64(0xfffff INLINE bits64 extractFloatx80Frac( floatx80 a ) { - return a.low; + return a.low; } @@ -65,7 +65,7 @@ INLINE bits64 extractFloatx80Frac( floatx80 a ) INLINE int32 extractFloatx80Exp( floatx80 a ) { - return a.high & 0x7FFF; + return a.high & 0x7FFF; } @@ -77,7 +77,7 @@ INLINE int32 extractFloatx80Exp( floatx80 a ) INLINE flag extractFloatx80Sign( floatx80 a ) { - return a.high>>15; + return a.high>>15; } @@ -89,12 +89,12 @@ INLINE flag extractFloatx80Sign( floatx80 a ) INLINE floatx80 propagateFloatx80NaNOneArg(floatx80 a) { - if (floatx80_is_signaling_nan(a)) - float_raise(float_flag_invalid); + if (floatx80_is_signaling_nan(a)) + float_raise(float_flag_invalid); - a.low |= U64(0xC000000000000000); + a.low |= U64(0xC000000000000000); - return a; + return a; } /*---------------------------------------------------------------------------- @@ -106,9 +106,9 @@ INLINE floatx80 propagateFloatx80NaNOneArg(floatx80 a) INLINE void normalizeFloatx80Subnormal(UINT64 aSig, INT32 *zExpPtr, UINT64 *zSigPtr) { - int shiftCount = countLeadingZeros64(aSig); - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; + int shiftCount = countLeadingZeros64(aSig); + *zSigPtr = aSig<<shiftCount; + *zExpPtr = 1 - shiftCount; } @@ -119,7 +119,7 @@ INLINE void normalizeFloatx80Subnormal(UINT64 aSig, INT32 *zExpPtr, UINT64 *zSig INLINE int floatx80_is_nan(floatx80 a) { - return ((a.high & 0x7FFF) == 0x7FFF) && (INT64) (a.low<<1); + return ((a.high & 0x7FFF) == 0x7FFF) && (INT64) (a.low<<1); } /*---------------------------------------------------------------------------- @@ -130,38 +130,38 @@ INLINE int floatx80_is_nan(floatx80 a) static floatx80 propagateFloatx80NaN(floatx80 a, floatx80 b) { - int aIsNaN = floatx80_is_nan(a); - int aIsSignalingNaN = floatx80_is_signaling_nan(a); - int bIsNaN = floatx80_is_nan(b); - int bIsSignalingNaN = floatx80_is_signaling_nan(b); - a.low |= U64(0xC000000000000000); - b.low |= U64(0xC000000000000000); - if (aIsSignalingNaN | bIsSignalingNaN) float_raise(float_flag_invalid); - if (aIsSignalingNaN) { - if (bIsSignalingNaN) goto returnLargerSignificand; - return bIsNaN ? b : a; - } - else if (aIsNaN) { - if (bIsSignalingNaN | ! bIsNaN) return a; - returnLargerSignificand: - if (a.low < b.low) return b; - if (b.low < a.low) return a; - return (a.high < b.high) ? a : b; - } - else { - return b; - } + int aIsNaN = floatx80_is_nan(a); + int aIsSignalingNaN = floatx80_is_signaling_nan(a); + int bIsNaN = floatx80_is_nan(b); + int bIsSignalingNaN = floatx80_is_signaling_nan(b); + a.low |= U64(0xC000000000000000); + b.low |= U64(0xC000000000000000); + if (aIsSignalingNaN | bIsSignalingNaN) float_raise(float_flag_invalid); + if (aIsSignalingNaN) { + if (bIsSignalingNaN) goto returnLargerSignificand; + return bIsNaN ? b : a; + } + else if (aIsNaN) { + if (bIsSignalingNaN | ! bIsNaN) return a; + returnLargerSignificand: + if (a.low < b.low) return b; + if (b.low < a.low) return a; + return (a.high < b.high) ? a : b; + } + else { + return b; + } } static const float128 float128_one = - packFloat_128(U64(0x3fff000000000000), U64(0x0000000000000000)); + packFloat_128(U64(0x3fff000000000000), U64(0x0000000000000000)); static const float128 float128_two = - packFloat_128(U64(0x4000000000000000), U64(0x0000000000000000)); + packFloat_128(U64(0x4000000000000000), U64(0x0000000000000000)); static const float128 float128_ln2inv2 = - packFloat_128(U64(0x400071547652b82f), U64(0xe1777d0ffda0d23a)); + packFloat_128(U64(0x400071547652b82f), U64(0xe1777d0ffda0d23a)); -#define SQRT2_HALF_SIG U64(0xb504f333f9de6484) +#define SQRT2_HALF_SIG U64(0xb504f333f9de6484) extern float128 OddPoly(float128 x, float128 *arr, unsigned n); @@ -169,15 +169,15 @@ extern float128 OddPoly(float128 x, float128 *arr, unsigned n); static float128 ln_arr[L2_ARR_SIZE] = { - PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 1 */ - PACK_FLOAT_128(0x3ffd555555555555, 0x5555555555555555), /* 3 */ - PACK_FLOAT_128(0x3ffc999999999999, 0x999999999999999a), /* 5 */ - PACK_FLOAT_128(0x3ffc249249249249, 0x2492492492492492), /* 7 */ - PACK_FLOAT_128(0x3ffbc71c71c71c71, 0xc71c71c71c71c71c), /* 9 */ - PACK_FLOAT_128(0x3ffb745d1745d174, 0x5d1745d1745d1746), /* 11 */ - PACK_FLOAT_128(0x3ffb3b13b13b13b1, 0x3b13b13b13b13b14), /* 13 */ - PACK_FLOAT_128(0x3ffb111111111111, 0x1111111111111111), /* 15 */ - PACK_FLOAT_128(0x3ffae1e1e1e1e1e1, 0xe1e1e1e1e1e1e1e2) /* 17 */ + PACK_FLOAT_128(0x3fff000000000000, 0x0000000000000000), /* 1 */ + PACK_FLOAT_128(0x3ffd555555555555, 0x5555555555555555), /* 3 */ + PACK_FLOAT_128(0x3ffc999999999999, 0x999999999999999a), /* 5 */ + PACK_FLOAT_128(0x3ffc249249249249, 0x2492492492492492), /* 7 */ + PACK_FLOAT_128(0x3ffbc71c71c71c71, 0xc71c71c71c71c71c), /* 9 */ + PACK_FLOAT_128(0x3ffb745d1745d174, 0x5d1745d1745d1746), /* 11 */ + PACK_FLOAT_128(0x3ffb3b13b13b13b1, 0x3b13b13b13b13b14), /* 13 */ + PACK_FLOAT_128(0x3ffb111111111111, 0x1111111111111111), /* 15 */ + PACK_FLOAT_128(0x3ffae1e1e1e1e1e1, 0xe1e1e1e1e1e1e1e2) /* 17 */ }; static float128 poly_ln(float128 x1) @@ -205,29 +205,29 @@ static float128 poly_ln(float128 x1) // 1-u // */ - return OddPoly(x1, ln_arr, L2_ARR_SIZE); + return OddPoly(x1, ln_arr, L2_ARR_SIZE); } /* required sqrt(2)/2 < x < sqrt(2) */ static float128 poly_l2(float128 x) { - /* using float128 for approximation */ - float128 x_p1 = float128_add(x, float128_one); - float128 x_m1 = float128_sub(x, float128_one); - x = float128_div(x_m1, x_p1); - x = poly_ln(x); - x = float128_mul(x, float128_ln2inv2); - return x; + /* using float128 for approximation */ + float128 x_p1 = float128_add(x, float128_one); + float128 x_m1 = float128_sub(x, float128_one); + x = float128_div(x_m1, x_p1); + x = poly_ln(x); + x = float128_mul(x, float128_ln2inv2); + return x; } static float128 poly_l2p1(float128 x) { - /* using float128 for approximation */ - float128 x_p2 = float128_add(x, float128_two); - x = float128_div(x, x_p2); - x = poly_ln(x); - x = float128_mul(x, float128_ln2inv2); - return x; + /* using float128 for approximation */ + float128 x_p2 = float128_add(x, float128_two); + x = float128_div(x, x_p2); + x = poly_ln(x); + x = float128_mul(x, float128_ln2inv2); + return x; } // ================================================= @@ -257,88 +257,88 @@ static float128 poly_l2p1(float128 x) static floatx80 fyl2x(floatx80 a, floatx80 b) { - UINT64 aSig = extractFloatx80Frac(a); - INT32 aExp = extractFloatx80Exp(a); - int aSign = extractFloatx80Sign(a); - UINT64 bSig = extractFloatx80Frac(b); - INT32 bExp = extractFloatx80Exp(b); - int bSign = extractFloatx80Sign(b); - - int zSign = bSign ^ 1; - - if (aExp == 0x7FFF) { - if ((UINT64) (aSig<<1) - || ((bExp == 0x7FFF) && (UINT64) (bSig<<1))) - { - return propagateFloatx80NaN(a, b); - } - if (aSign) - { + UINT64 aSig = extractFloatx80Frac(a); + INT32 aExp = extractFloatx80Exp(a); + int aSign = extractFloatx80Sign(a); + UINT64 bSig = extractFloatx80Frac(b); + INT32 bExp = extractFloatx80Exp(b); + int bSign = extractFloatx80Sign(b); + + int zSign = bSign ^ 1; + + if (aExp == 0x7FFF) { + if ((UINT64) (aSig<<1) + || ((bExp == 0x7FFF) && (UINT64) (bSig<<1))) + { + return propagateFloatx80NaN(a, b); + } + if (aSign) + { invalid: float_raise(float_flag_invalid); return floatx80_default_nan; - } - else { - if (bExp == 0) { - if (bSig == 0) goto invalid; - float_raise(float_flag_denormal); - } - return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); - } - } - if (bExp == 0x7FFF) - { - if ((UINT64) (bSig<<1)) return propagateFloatx80NaN(a, b); - if (aSign && (UINT64)(aExp | aSig)) goto invalid; - if (aSig && (aExp == 0)) - float_raise(float_flag_denormal); - if (aExp < 0x3FFF) { - return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); - } - if (aExp == 0x3FFF && ((UINT64) (aSig<<1) == 0)) goto invalid; - return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); - } - if (aExp == 0) { - if (aSig == 0) { - if ((bExp | bSig) == 0) goto invalid; - float_raise(float_flag_divbyzero); - return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); - } - if (aSign) goto invalid; - float_raise(float_flag_denormal); - normalizeFloatx80Subnormal(aSig, &aExp, &aSig); - } - if (aSign) goto invalid; - if (bExp == 0) { - if (bSig == 0) { - if (aExp < 0x3FFF) return packFloatx80(zSign, 0, 0); - return packFloatx80(bSign, 0, 0); - } - float_raise(float_flag_denormal); - normalizeFloatx80Subnormal(bSig, &bExp, &bSig); - } - if (aExp == 0x3FFF && ((UINT64) (aSig<<1) == 0)) - return packFloatx80(bSign, 0, 0); - - float_raise(float_flag_inexact); - - int ExpDiff = aExp - 0x3FFF; - aExp = 0; - if (aSig >= SQRT2_HALF_SIG) { - ExpDiff++; - aExp--; - } - - /* ******************************** */ - /* using float128 for approximation */ - /* ******************************** */ - - UINT64 zSig0, zSig1; - shift128Right(aSig<<1, 0, 16, &zSig0, &zSig1); - float128 x = packFloat128(0, aExp+0x3FFF, zSig0, zSig1); - x = poly_l2(x); - x = float128_add(x, int64_to_float128((INT64) ExpDiff)); - return floatx80_mul(b, float128_to_floatx80(x)); + } + else { + if (bExp == 0) { + if (bSig == 0) goto invalid; + float_raise(float_flag_denormal); + } + return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); + } + } + if (bExp == 0x7FFF) + { + if ((UINT64) (bSig<<1)) return propagateFloatx80NaN(a, b); + if (aSign && (UINT64)(aExp | aSig)) goto invalid; + if (aSig && (aExp == 0)) + float_raise(float_flag_denormal); + if (aExp < 0x3FFF) { + return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); + } + if (aExp == 0x3FFF && ((UINT64) (aSig<<1) == 0)) goto invalid; + return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); + } + if (aExp == 0) { + if (aSig == 0) { + if ((bExp | bSig) == 0) goto invalid; + float_raise(float_flag_divbyzero); + return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); + } + if (aSign) goto invalid; + float_raise(float_flag_denormal); + normalizeFloatx80Subnormal(aSig, &aExp, &aSig); + } + if (aSign) goto invalid; + if (bExp == 0) { + if (bSig == 0) { + if (aExp < 0x3FFF) return packFloatx80(zSign, 0, 0); + return packFloatx80(bSign, 0, 0); + } + float_raise(float_flag_denormal); + normalizeFloatx80Subnormal(bSig, &bExp, &bSig); + } + if (aExp == 0x3FFF && ((UINT64) (aSig<<1) == 0)) + return packFloatx80(bSign, 0, 0); + + float_raise(float_flag_inexact); + + int ExpDiff = aExp - 0x3FFF; + aExp = 0; + if (aSig >= SQRT2_HALF_SIG) { + ExpDiff++; + aExp--; + } + + /* ******************************** */ + /* using float128 for approximation */ + /* ******************************** */ + + UINT64 zSig0, zSig1; + shift128Right(aSig<<1, 0, 16, &zSig0, &zSig1); + float128 x = packFloat128(0, aExp+0x3FFF, zSig0, zSig1); + x = poly_l2(x); + x = float128_add(x, int64_to_float128((INT64) ExpDiff)); + return floatx80_mul(b, float128_to_floatx80(x)); } // ================================================= @@ -368,123 +368,123 @@ invalid: floatx80 fyl2xp1(floatx80 a, floatx80 b) { - INT32 aExp, bExp; - UINT64 aSig, bSig, zSig0, zSig1, zSig2; - int aSign, bSign; - - aSig = extractFloatx80Frac(a); - aExp = extractFloatx80Exp(a); - aSign = extractFloatx80Sign(a); - bSig = extractFloatx80Frac(b); - bExp = extractFloatx80Exp(b); - bSign = extractFloatx80Sign(b); - int zSign = aSign ^ bSign; - - if (aExp == 0x7FFF) { - if ((UINT64) (aSig<<1) - || ((bExp == 0x7FFF) && (UINT64) (bSig<<1))) - { - return propagateFloatx80NaN(a, b); - } - if (aSign) - { + INT32 aExp, bExp; + UINT64 aSig, bSig, zSig0, zSig1, zSig2; + int aSign, bSign; + + aSig = extractFloatx80Frac(a); + aExp = extractFloatx80Exp(a); + aSign = extractFloatx80Sign(a); + bSig = extractFloatx80Frac(b); + bExp = extractFloatx80Exp(b); + bSign = extractFloatx80Sign(b); + int zSign = aSign ^ bSign; + + if (aExp == 0x7FFF) { + if ((UINT64) (aSig<<1) + || ((bExp == 0x7FFF) && (UINT64) (bSig<<1))) + { + return propagateFloatx80NaN(a, b); + } + if (aSign) + { invalid: float_raise(float_flag_invalid); return floatx80_default_nan; - } - else { - if (bExp == 0) { - if (bSig == 0) goto invalid; - float_raise(float_flag_denormal); - } - return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); - } - } - if (bExp == 0x7FFF) - { - if ((UINT64) (bSig<<1)) - return propagateFloatx80NaN(a, b); - - if (aExp == 0) { - if (aSig == 0) goto invalid; - float_raise(float_flag_denormal); - } - - return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); - } - if (aExp == 0) { - if (aSig == 0) { - if (bSig && (bExp == 0)) float_raise(float_flag_denormal); - return packFloatx80(zSign, 0, 0); - } - float_raise(float_flag_denormal); - normalizeFloatx80Subnormal(aSig, &aExp, &aSig); - } - if (bExp == 0) { - if (bSig == 0) return packFloatx80(zSign, 0, 0); - float_raise(float_flag_denormal); - normalizeFloatx80Subnormal(bSig, &bExp, &bSig); - } - - float_raise(float_flag_inexact); - - if (aSign && aExp >= 0x3FFF) - return a; - - if (aExp >= 0x3FFC) // big argument - { - return fyl2x(floatx80_add(a, floatx80_one), b); - } - - // handle tiny argument - if (aExp < EXP_BIAS-70) - { - // first order approximation, return (a*b)/ln(2) - INT32 zExp = aExp + FLOAT_LN2INV_EXP - 0x3FFE; + } + else { + if (bExp == 0) { + if (bSig == 0) goto invalid; + float_raise(float_flag_denormal); + } + return packFloatx80(bSign, 0x7FFF, U64(0x8000000000000000)); + } + } + if (bExp == 0x7FFF) + { + if ((UINT64) (bSig<<1)) + return propagateFloatx80NaN(a, b); + + if (aExp == 0) { + if (aSig == 0) goto invalid; + float_raise(float_flag_denormal); + } + + return packFloatx80(zSign, 0x7FFF, U64(0x8000000000000000)); + } + if (aExp == 0) { + if (aSig == 0) { + if (bSig && (bExp == 0)) float_raise(float_flag_denormal); + return packFloatx80(zSign, 0, 0); + } + float_raise(float_flag_denormal); + normalizeFloatx80Subnormal(aSig, &aExp, &aSig); + } + if (bExp == 0) { + if (bSig == 0) return packFloatx80(zSign, 0, 0); + float_raise(float_flag_denormal); + normalizeFloatx80Subnormal(bSig, &bExp, &bSig); + } + + float_raise(float_flag_inexact); + + if (aSign && aExp >= 0x3FFF) + return a; + + if (aExp >= 0x3FFC) // big argument + { + return fyl2x(floatx80_add(a, floatx80_one), b); + } + + // handle tiny argument + if (aExp < EXP_BIAS-70) + { + // first order approximation, return (a*b)/ln(2) + INT32 zExp = aExp + FLOAT_LN2INV_EXP - 0x3FFE; mul128By64To192(FLOAT_LN2INV_HI, FLOAT_LN2INV_LO, aSig, &zSig0, &zSig1, &zSig2); - if (0 < (INT64) zSig0) { - shortShift128Left(zSig0, zSig1, 1, &zSig0, &zSig1); - --zExp; - } + if (0 < (INT64) zSig0) { + shortShift128Left(zSig0, zSig1, 1, &zSig0, &zSig1); + --zExp; + } - zExp = zExp + bExp - 0x3FFE; + zExp = zExp + bExp - 0x3FFE; mul128By64To192(zSig0, zSig1, bSig, &zSig0, &zSig1, &zSig2); - if (0 < (INT64) zSig0) { - shortShift128Left(zSig0, zSig1, 1, &zSig0, &zSig1); - --zExp; - } - - return - roundAndPackFloatx80(80, aSign ^ bSign, zExp, zSig0, zSig1); - } - - /* ******************************** */ - /* using float128 for approximation */ - /* ******************************** */ - - shift128Right(aSig<<1, 0, 16, &zSig0, &zSig1); - float128 x = packFloat128(aSign, aExp, zSig0, zSig1); - x = poly_l2p1(x); - return floatx80_mul(b, float128_to_floatx80(x)); + if (0 < (INT64) zSig0) { + shortShift128Left(zSig0, zSig1, 1, &zSig0, &zSig1); + --zExp; + } + + return + roundAndPackFloatx80(80, aSign ^ bSign, zExp, zSig0, zSig1); + } + + /* ******************************** */ + /* using float128 for approximation */ + /* ******************************** */ + + shift128Right(aSig<<1, 0, 16, &zSig0, &zSig1); + float128 x = packFloat128(aSign, aExp, zSig0, zSig1); + x = poly_l2p1(x); + return floatx80_mul(b, float128_to_floatx80(x)); } floatx80 floatx80_flognp1(floatx80 a) { - return fyl2xp1(a, floatx80_ln_2); + return fyl2xp1(a, floatx80_ln_2); } floatx80 floatx80_flogn(floatx80 a) { - return fyl2x(a, floatx80_ln_2); + return fyl2x(a, floatx80_ln_2); } floatx80 floatx80_flog2(floatx80 a) { - return fyl2x(a, floatx80_one); + return fyl2x(a, floatx80_one); } floatx80 floatx80_flog10(floatx80 a) { - return fyl2x(a, floatx80_log10_2); + return fyl2x(a, floatx80_log10_2); } diff --git a/src/lib/softfloat/softfloat.c b/src/lib/softfloat/softfloat.c index fe0137787bc..89a191aa681 100644 --- a/src/lib/softfloat/softfloat.c +++ b/src/lib/softfloat/softfloat.c @@ -67,39 +67,39 @@ int8 float_rounding_mode = float_round_nearest_even; static int32 roundAndPackInt32( flag zSign, bits64 absZ ) { - int8 roundingMode; - flag roundNearestEven; - int8 roundIncrement, roundBits; - int32 z; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x40; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x7F; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = absZ & 0x7F; - absZ = ( absZ + roundIncrement )>>7; - absZ &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); - z = absZ; - if ( zSign ) z = - z; - if ( ( absZ>>32 ) || ( z && ( ( z < 0 ) ^ zSign ) ) ) { - float_raise( float_flag_invalid ); - return zSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - return z; + int8 roundingMode; + flag roundNearestEven; + int8 roundIncrement, roundBits; + int32 z; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + roundIncrement = 0x40; + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + roundIncrement = 0; + } + else { + roundIncrement = 0x7F; + if ( zSign ) { + if ( roundingMode == float_round_up ) roundIncrement = 0; + } + else { + if ( roundingMode == float_round_down ) roundIncrement = 0; + } + } + } + roundBits = absZ & 0x7F; + absZ = ( absZ + roundIncrement )>>7; + absZ &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); + z = absZ; + if ( zSign ) z = - z; + if ( ( absZ>>32 ) || ( z && ( ( z < 0 ) ^ zSign ) ) ) { + float_raise( float_flag_invalid ); + return zSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; + } + if ( roundBits ) float_exception_flags |= float_flag_inexact; + return z; } @@ -117,42 +117,42 @@ static int32 roundAndPackInt32( flag zSign, bits64 absZ ) static int64 roundAndPackInt64( flag zSign, bits64 absZ0, bits64 absZ1 ) { - int8 roundingMode; - flag roundNearestEven, increment; - int64 z; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - increment = ( (sbits64) absZ1 < 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - increment = 0; - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && absZ1; - } - else { - increment = ( roundingMode == float_round_up ) && absZ1; - } - } - } - if ( increment ) { - ++absZ0; - if ( absZ0 == 0 ) goto overflow; - absZ0 &= ~ ( ( (bits64) ( absZ1<<1 ) == 0 ) & roundNearestEven ); - } - z = absZ0; - if ( zSign ) z = - z; - if ( z && ( ( z < 0 ) ^ zSign ) ) { - overflow: - float_raise( float_flag_invalid ); - return - zSign ? (sbits64) LIT64( 0x8000000000000000 ) - : LIT64( 0x7FFFFFFFFFFFFFFF ); - } - if ( absZ1 ) float_exception_flags |= float_flag_inexact; - return z; + int8 roundingMode; + flag roundNearestEven, increment; + int64 z; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + increment = ( (sbits64) absZ1 < 0 ); + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + increment = 0; + } + else { + if ( zSign ) { + increment = ( roundingMode == float_round_down ) && absZ1; + } + else { + increment = ( roundingMode == float_round_up ) && absZ1; + } + } + } + if ( increment ) { + ++absZ0; + if ( absZ0 == 0 ) goto overflow; + absZ0 &= ~ ( ( (bits64) ( absZ1<<1 ) == 0 ) & roundNearestEven ); + } + z = absZ0; + if ( zSign ) z = - z; + if ( z && ( ( z < 0 ) ^ zSign ) ) { + overflow: + float_raise( float_flag_invalid ); + return + zSign ? (sbits64) LIT64( 0x8000000000000000 ) + : LIT64( 0x7FFFFFFFFFFFFFFF ); + } + if ( absZ1 ) float_exception_flags |= float_flag_inexact; + return z; } @@ -163,7 +163,7 @@ static int64 roundAndPackInt64( flag zSign, bits64 absZ0, bits64 absZ1 ) INLINE bits32 extractFloat32Frac( float32 a ) { - return a & 0x007FFFFF; + return a & 0x007FFFFF; } @@ -174,7 +174,7 @@ INLINE bits32 extractFloat32Frac( float32 a ) INLINE int16 extractFloat32Exp( float32 a ) { - return ( a>>23 ) & 0xFF; + return ( a>>23 ) & 0xFF; } @@ -185,7 +185,7 @@ INLINE int16 extractFloat32Exp( float32 a ) INLINE flag extractFloat32Sign( float32 a ) { - return a>>31; + return a>>31; } @@ -197,13 +197,13 @@ INLINE flag extractFloat32Sign( float32 a ) *----------------------------------------------------------------------------*/ static void - normalizeFloat32Subnormal( bits32 aSig, int16 *zExpPtr, bits32 *zSigPtr ) + normalizeFloat32Subnormal( bits32 aSig, int16 *zExpPtr, bits32 *zSigPtr ) { - int8 shiftCount; + int8 shiftCount; - shiftCount = countLeadingZeros32( aSig ) - 8; - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; + shiftCount = countLeadingZeros32( aSig ) - 8; + *zSigPtr = aSig<<shiftCount; + *zExpPtr = 1 - shiftCount; } @@ -221,7 +221,7 @@ static void INLINE float32 packFloat32( flag zSign, int16 zExp, bits32 zSig ) { - return ( ( (bits32) zSign )<<31 ) + ( ( (bits32) zExp )<<23 ) + zSig; + return ( ( (bits32) zSign )<<31 ) + ( ( (bits32) zExp )<<23 ) + zSig; } @@ -249,53 +249,53 @@ INLINE float32 packFloat32( flag zSign, int16 zExp, bits32 zSig ) static float32 roundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) { - int8 roundingMode; - flag roundNearestEven; - int8 roundIncrement, roundBits; - flag isTiny; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x40; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x7F; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig & 0x7F; - if ( 0xFD <= (bits16) zExp ) { - if ( ( 0xFD < zExp ) - || ( ( zExp == 0xFD ) - && ( (sbits32) ( zSig + roundIncrement ) < 0 ) ) - ) { - float_raise( float_flag_overflow | float_flag_inexact ); - return packFloat32( zSign, 0xFF, 0 ) - ( roundIncrement == 0 ); - } - if ( zExp < 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < -1 ) - || ( zSig + roundIncrement < 0x80000000 ); - shift32RightJamming( zSig, - zExp, &zSig ); - zExp = 0; - roundBits = zSig & 0x7F; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig = ( zSig + roundIncrement )>>7; - zSig &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); - if ( zSig == 0 ) zExp = 0; - return packFloat32( zSign, zExp, zSig ); + int8 roundingMode; + flag roundNearestEven; + int8 roundIncrement, roundBits; + flag isTiny; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + roundIncrement = 0x40; + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + roundIncrement = 0; + } + else { + roundIncrement = 0x7F; + if ( zSign ) { + if ( roundingMode == float_round_up ) roundIncrement = 0; + } + else { + if ( roundingMode == float_round_down ) roundIncrement = 0; + } + } + } + roundBits = zSig & 0x7F; + if ( 0xFD <= (bits16) zExp ) { + if ( ( 0xFD < zExp ) + || ( ( zExp == 0xFD ) + && ( (sbits32) ( zSig + roundIncrement ) < 0 ) ) + ) { + float_raise( float_flag_overflow | float_flag_inexact ); + return packFloat32( zSign, 0xFF, 0 ) - ( roundIncrement == 0 ); + } + if ( zExp < 0 ) { + isTiny = + ( float_detect_tininess == float_tininess_before_rounding ) + || ( zExp < -1 ) + || ( zSig + roundIncrement < 0x80000000 ); + shift32RightJamming( zSig, - zExp, &zSig ); + zExp = 0; + roundBits = zSig & 0x7F; + if ( isTiny && roundBits ) float_raise( float_flag_underflow ); + } + } + if ( roundBits ) float_exception_flags |= float_flag_inexact; + zSig = ( zSig + roundIncrement )>>7; + zSig &= ~ ( ( ( roundBits ^ 0x40 ) == 0 ) & roundNearestEven ); + if ( zSig == 0 ) zExp = 0; + return packFloat32( zSign, zExp, zSig ); } @@ -309,12 +309,12 @@ static float32 roundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) *----------------------------------------------------------------------------*/ static float32 - normalizeRoundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) + normalizeRoundAndPackFloat32( flag zSign, int16 zExp, bits32 zSig ) { - int8 shiftCount; + int8 shiftCount; - shiftCount = countLeadingZeros32( zSig ) - 1; - return roundAndPackFloat32( zSign, zExp - shiftCount, zSig<<shiftCount ); + shiftCount = countLeadingZeros32( zSig ) - 1; + return roundAndPackFloat32( zSign, zExp - shiftCount, zSig<<shiftCount ); } @@ -325,7 +325,7 @@ static float32 INLINE bits64 extractFloat64Frac( float64 a ) { - return a & LIT64( 0x000FFFFFFFFFFFFF ); + return a & LIT64( 0x000FFFFFFFFFFFFF ); } @@ -336,7 +336,7 @@ INLINE bits64 extractFloat64Frac( float64 a ) INLINE int16 extractFloat64Exp( float64 a ) { - return ( a>>52 ) & 0x7FF; + return ( a>>52 ) & 0x7FF; } @@ -347,7 +347,7 @@ INLINE int16 extractFloat64Exp( float64 a ) INLINE flag extractFloat64Sign( float64 a ) { - return a>>63; + return a>>63; } @@ -359,13 +359,13 @@ INLINE flag extractFloat64Sign( float64 a ) *----------------------------------------------------------------------------*/ static void - normalizeFloat64Subnormal( bits64 aSig, int16 *zExpPtr, bits64 *zSigPtr ) + normalizeFloat64Subnormal( bits64 aSig, int16 *zExpPtr, bits64 *zSigPtr ) { - int8 shiftCount; + int8 shiftCount; - shiftCount = countLeadingZeros64( aSig ) - 11; - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; + shiftCount = countLeadingZeros64( aSig ) - 11; + *zSigPtr = aSig<<shiftCount; + *zExpPtr = 1 - shiftCount; } @@ -383,7 +383,7 @@ static void INLINE float64 packFloat64( flag zSign, int16 zExp, bits64 zSig ) { - return ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<52 ) + zSig; + return ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<52 ) + zSig; } @@ -411,53 +411,53 @@ INLINE float64 packFloat64( flag zSign, int16 zExp, bits64 zSig ) static float64 roundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) { - int8 roundingMode; - flag roundNearestEven; - int16 roundIncrement, roundBits; - flag isTiny; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - roundIncrement = 0x200; - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = 0x3FF; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig & 0x3FF; - if ( 0x7FD <= (bits16) zExp ) { - if ( ( 0x7FD < zExp ) - || ( ( zExp == 0x7FD ) - && ( (sbits64) ( zSig + roundIncrement ) < 0 ) ) - ) { - float_raise( float_flag_overflow | float_flag_inexact ); - return packFloat64( zSign, 0x7FF, 0 ) - ( roundIncrement == 0 ); - } - if ( zExp < 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < -1 ) - || ( zSig + roundIncrement < LIT64( 0x8000000000000000 ) ); - shift64RightJamming( zSig, - zExp, &zSig ); - zExp = 0; - roundBits = zSig & 0x3FF; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig = ( zSig + roundIncrement )>>10; - zSig &= ~ ( ( ( roundBits ^ 0x200 ) == 0 ) & roundNearestEven ); - if ( zSig == 0 ) zExp = 0; - return packFloat64( zSign, zExp, zSig ); + int8 roundingMode; + flag roundNearestEven; + int16 roundIncrement, roundBits; + flag isTiny; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + roundIncrement = 0x200; + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + roundIncrement = 0; + } + else { + roundIncrement = 0x3FF; + if ( zSign ) { + if ( roundingMode == float_round_up ) roundIncrement = 0; + } + else { + if ( roundingMode == float_round_down ) roundIncrement = 0; + } + } + } + roundBits = zSig & 0x3FF; + if ( 0x7FD <= (bits16) zExp ) { + if ( ( 0x7FD < zExp ) + || ( ( zExp == 0x7FD ) + && ( (sbits64) ( zSig + roundIncrement ) < 0 ) ) + ) { + float_raise( float_flag_overflow | float_flag_inexact ); + return packFloat64( zSign, 0x7FF, 0 ) - ( roundIncrement == 0 ); + } + if ( zExp < 0 ) { + isTiny = + ( float_detect_tininess == float_tininess_before_rounding ) + || ( zExp < -1 ) + || ( zSig + roundIncrement < LIT64( 0x8000000000000000 ) ); + shift64RightJamming( zSig, - zExp, &zSig ); + zExp = 0; + roundBits = zSig & 0x3FF; + if ( isTiny && roundBits ) float_raise( float_flag_underflow ); + } + } + if ( roundBits ) float_exception_flags |= float_flag_inexact; + zSig = ( zSig + roundIncrement )>>10; + zSig &= ~ ( ( ( roundBits ^ 0x200 ) == 0 ) & roundNearestEven ); + if ( zSig == 0 ) zExp = 0; + return packFloat64( zSign, zExp, zSig ); } @@ -471,12 +471,12 @@ static float64 roundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) *----------------------------------------------------------------------------*/ static float64 - normalizeRoundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) + normalizeRoundAndPackFloat64( flag zSign, int16 zExp, bits64 zSig ) { - int8 shiftCount; + int8 shiftCount; - shiftCount = countLeadingZeros64( zSig ) - 1; - return roundAndPackFloat64( zSign, zExp - shiftCount, zSig<<shiftCount ); + shiftCount = countLeadingZeros64( zSig ) - 1; + return roundAndPackFloat64( zSign, zExp - shiftCount, zSig<<shiftCount ); } @@ -490,13 +490,13 @@ static float64 *----------------------------------------------------------------------------*/ static void - normalizeFloatx80Subnormal( bits64 aSig, int32 *zExpPtr, bits64 *zSigPtr ) + normalizeFloatx80Subnormal( bits64 aSig, int32 *zExpPtr, bits64 *zSigPtr ) { - int8 shiftCount; + int8 shiftCount; - shiftCount = countLeadingZeros64( aSig ); - *zSigPtr = aSig<<shiftCount; - *zExpPtr = 1 - shiftCount; + shiftCount = countLeadingZeros64( aSig ); + *zSigPtr = aSig<<shiftCount; + *zExpPtr = 1 - shiftCount; } @@ -527,161 +527,161 @@ static void // roundAndPackFloatx80 is now also used in fyl2x.c /* static */ floatx80 - roundAndPackFloatx80( - int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 - ) + roundAndPackFloatx80( + int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 + ) { - int8 roundingMode; - flag roundNearestEven, increment, isTiny; - int64 roundIncrement, roundMask, roundBits; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - if ( roundingPrecision == 80 ) goto precision80; - if ( roundingPrecision == 64 ) { - roundIncrement = LIT64( 0x0000000000000400 ); - roundMask = LIT64( 0x00000000000007FF ); - } - else if ( roundingPrecision == 32 ) { - roundIncrement = LIT64( 0x0000008000000000 ); - roundMask = LIT64( 0x000000FFFFFFFFFF ); - } - else { - goto precision80; - } - zSig0 |= ( zSig1 != 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - roundIncrement = 0; - } - else { - roundIncrement = roundMask; - if ( zSign ) { - if ( roundingMode == float_round_up ) roundIncrement = 0; - } - else { - if ( roundingMode == float_round_down ) roundIncrement = 0; - } - } - } - roundBits = zSig0 & roundMask; - if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { - if ( ( 0x7FFE < zExp ) - || ( ( zExp == 0x7FFE ) && ( zSig0 + roundIncrement < zSig0 ) ) - ) { - goto overflow; - } - if ( zExp <= 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < 0 ) - || ( zSig0 <= zSig0 + roundIncrement ); - shift64RightJamming( zSig0, 1 - zExp, &zSig0 ); - zExp = 0; - roundBits = zSig0 & roundMask; - if ( isTiny && roundBits ) float_raise( float_flag_underflow ); - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig0 += roundIncrement; - if ( (sbits64) zSig0 < 0 ) zExp = 1; - roundIncrement = roundMask + 1; - if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { - roundMask |= roundIncrement; - } - zSig0 &= ~ roundMask; - return packFloatx80( zSign, zExp, zSig0 ); - } - } - if ( roundBits ) float_exception_flags |= float_flag_inexact; - zSig0 += roundIncrement; - if ( zSig0 < roundIncrement ) { - ++zExp; - zSig0 = LIT64( 0x8000000000000000 ); - } - roundIncrement = roundMask + 1; - if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { - roundMask |= roundIncrement; - } - zSig0 &= ~ roundMask; - if ( zSig0 == 0 ) zExp = 0; - return packFloatx80( zSign, zExp, zSig0 ); - precision80: - increment = ( (sbits64) zSig1 < 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - increment = 0; - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig1; - } - else { - increment = ( roundingMode == float_round_up ) && zSig1; - } - } - } - if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { - if ( ( 0x7FFE < zExp ) - || ( ( zExp == 0x7FFE ) - && ( zSig0 == LIT64( 0xFFFFFFFFFFFFFFFF ) ) - && increment - ) - ) { - roundMask = 0; - overflow: - float_raise( float_flag_overflow | float_flag_inexact ); - if ( ( roundingMode == float_round_to_zero ) - || ( zSign && ( roundingMode == float_round_up ) ) - || ( ! zSign && ( roundingMode == float_round_down ) ) - ) { - return packFloatx80( zSign, 0x7FFE, ~ roundMask ); - } - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( zExp <= 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < 0 ) - || ! increment - || ( zSig0 < LIT64( 0xFFFFFFFFFFFFFFFF ) ); - shift64ExtraRightJamming( zSig0, zSig1, 1 - zExp, &zSig0, &zSig1 ); - zExp = 0; - if ( isTiny && zSig1 ) float_raise( float_flag_underflow ); - if ( zSig1 ) float_exception_flags |= float_flag_inexact; - if ( roundNearestEven ) { - increment = ( (sbits64) zSig1 < 0 ); - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig1; - } - else { - increment = ( roundingMode == float_round_up ) && zSig1; - } - } - if ( increment ) { - ++zSig0; - zSig0 &= - ~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); - if ( (sbits64) zSig0 < 0 ) zExp = 1; - } - return packFloatx80( zSign, zExp, zSig0 ); - } - } - if ( zSig1 ) float_exception_flags |= float_flag_inexact; - if ( increment ) { - ++zSig0; - if ( zSig0 == 0 ) { - ++zExp; - zSig0 = LIT64( 0x8000000000000000 ); - } - else { - zSig0 &= ~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); - } - } - else { - if ( zSig0 == 0 ) zExp = 0; - } - return packFloatx80( zSign, zExp, zSig0 ); + int8 roundingMode; + flag roundNearestEven, increment, isTiny; + int64 roundIncrement, roundMask, roundBits; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + if ( roundingPrecision == 80 ) goto precision80; + if ( roundingPrecision == 64 ) { + roundIncrement = LIT64( 0x0000000000000400 ); + roundMask = LIT64( 0x00000000000007FF ); + } + else if ( roundingPrecision == 32 ) { + roundIncrement = LIT64( 0x0000008000000000 ); + roundMask = LIT64( 0x000000FFFFFFFFFF ); + } + else { + goto precision80; + } + zSig0 |= ( zSig1 != 0 ); + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + roundIncrement = 0; + } + else { + roundIncrement = roundMask; + if ( zSign ) { + if ( roundingMode == float_round_up ) roundIncrement = 0; + } + else { + if ( roundingMode == float_round_down ) roundIncrement = 0; + } + } + } + roundBits = zSig0 & roundMask; + if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { + if ( ( 0x7FFE < zExp ) + || ( ( zExp == 0x7FFE ) && ( zSig0 + roundIncrement < zSig0 ) ) + ) { + goto overflow; + } + if ( zExp <= 0 ) { + isTiny = + ( float_detect_tininess == float_tininess_before_rounding ) + || ( zExp < 0 ) + || ( zSig0 <= zSig0 + roundIncrement ); + shift64RightJamming( zSig0, 1 - zExp, &zSig0 ); + zExp = 0; + roundBits = zSig0 & roundMask; + if ( isTiny && roundBits ) float_raise( float_flag_underflow ); + if ( roundBits ) float_exception_flags |= float_flag_inexact; + zSig0 += roundIncrement; + if ( (sbits64) zSig0 < 0 ) zExp = 1; + roundIncrement = roundMask + 1; + if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { + roundMask |= roundIncrement; + } + zSig0 &= ~ roundMask; + return packFloatx80( zSign, zExp, zSig0 ); + } + } + if ( roundBits ) float_exception_flags |= float_flag_inexact; + zSig0 += roundIncrement; + if ( zSig0 < roundIncrement ) { + ++zExp; + zSig0 = LIT64( 0x8000000000000000 ); + } + roundIncrement = roundMask + 1; + if ( roundNearestEven && ( roundBits<<1 == roundIncrement ) ) { + roundMask |= roundIncrement; + } + zSig0 &= ~ roundMask; + if ( zSig0 == 0 ) zExp = 0; + return packFloatx80( zSign, zExp, zSig0 ); + precision80: + increment = ( (sbits64) zSig1 < 0 ); + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + increment = 0; + } + else { + if ( zSign ) { + increment = ( roundingMode == float_round_down ) && zSig1; + } + else { + increment = ( roundingMode == float_round_up ) && zSig1; + } + } + } + if ( 0x7FFD <= (bits32) ( zExp - 1 ) ) { + if ( ( 0x7FFE < zExp ) + || ( ( zExp == 0x7FFE ) + && ( zSig0 == LIT64( 0xFFFFFFFFFFFFFFFF ) ) + && increment + ) + ) { + roundMask = 0; + overflow: + float_raise( float_flag_overflow | float_flag_inexact ); + if ( ( roundingMode == float_round_to_zero ) + || ( zSign && ( roundingMode == float_round_up ) ) + || ( ! zSign && ( roundingMode == float_round_down ) ) + ) { + return packFloatx80( zSign, 0x7FFE, ~ roundMask ); + } + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( zExp <= 0 ) { + isTiny = + ( float_detect_tininess == float_tininess_before_rounding ) + || ( zExp < 0 ) + || ! increment + || ( zSig0 < LIT64( 0xFFFFFFFFFFFFFFFF ) ); + shift64ExtraRightJamming( zSig0, zSig1, 1 - zExp, &zSig0, &zSig1 ); + zExp = 0; + if ( isTiny && zSig1 ) float_raise( float_flag_underflow ); + if ( zSig1 ) float_exception_flags |= float_flag_inexact; + if ( roundNearestEven ) { + increment = ( (sbits64) zSig1 < 0 ); + } + else { + if ( zSign ) { + increment = ( roundingMode == float_round_down ) && zSig1; + } + else { + increment = ( roundingMode == float_round_up ) && zSig1; + } + } + if ( increment ) { + ++zSig0; + zSig0 &= + ~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); + if ( (sbits64) zSig0 < 0 ) zExp = 1; + } + return packFloatx80( zSign, zExp, zSig0 ); + } + } + if ( zSig1 ) float_exception_flags |= float_flag_inexact; + if ( increment ) { + ++zSig0; + if ( zSig0 == 0 ) { + ++zExp; + zSig0 = LIT64( 0x8000000000000000 ); + } + else { + zSig0 &= ~ ( ( (bits64) ( zSig1<<1 ) == 0 ) & roundNearestEven ); + } + } + else { + if ( zSig0 == 0 ) zExp = 0; + } + return packFloatx80( zSign, zExp, zSig0 ); } @@ -695,22 +695,22 @@ static void *----------------------------------------------------------------------------*/ static floatx80 - normalizeRoundAndPackFloatx80( - int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 - ) + normalizeRoundAndPackFloatx80( + int8 roundingPrecision, flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 + ) { - int8 shiftCount; - - if ( zSig0 == 0 ) { - zSig0 = zSig1; - zSig1 = 0; - zExp -= 64; - } - shiftCount = countLeadingZeros64( zSig0 ); - shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); - zExp -= shiftCount; - return - roundAndPackFloatx80( roundingPrecision, zSign, zExp, zSig0, zSig1 ); + int8 shiftCount; + + if ( zSig0 == 0 ) { + zSig0 = zSig1; + zSig1 = 0; + zExp -= 64; + } + shiftCount = countLeadingZeros64( zSig0 ); + shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); + zExp -= shiftCount; + return + roundAndPackFloatx80( roundingPrecision, zSign, zExp, zSig0, zSig1 ); } @@ -726,7 +726,7 @@ static floatx80 INLINE bits64 extractFloat128Frac1( float128 a ) { - return a.low; + return a.low; } @@ -738,7 +738,7 @@ INLINE bits64 extractFloat128Frac1( float128 a ) INLINE bits64 extractFloat128Frac0( float128 a ) { - return a.high & LIT64( 0x0000FFFFFFFFFFFF ); + return a.high & LIT64( 0x0000FFFFFFFFFFFF ); } @@ -750,7 +750,7 @@ INLINE bits64 extractFloat128Frac0( float128 a ) INLINE int32 extractFloat128Exp( float128 a ) { - return ( a.high>>48 ) & 0x7FFF; + return ( a.high>>48 ) & 0x7FFF; } @@ -761,7 +761,7 @@ INLINE int32 extractFloat128Exp( float128 a ) INLINE flag extractFloat128Sign( float128 a ) { - return a.high>>63; + return a.high>>63; } @@ -776,33 +776,33 @@ INLINE flag extractFloat128Sign( float128 a ) *----------------------------------------------------------------------------*/ static void - normalizeFloat128Subnormal( - bits64 aSig0, - bits64 aSig1, - int32 *zExpPtr, - bits64 *zSig0Ptr, - bits64 *zSig1Ptr - ) + normalizeFloat128Subnormal( + bits64 aSig0, + bits64 aSig1, + int32 *zExpPtr, + bits64 *zSig0Ptr, + bits64 *zSig1Ptr + ) { - int8 shiftCount; - - if ( aSig0 == 0 ) { - shiftCount = countLeadingZeros64( aSig1 ) - 15; - if ( shiftCount < 0 ) { - *zSig0Ptr = aSig1>>( - shiftCount ); - *zSig1Ptr = aSig1<<( shiftCount & 63 ); - } - else { - *zSig0Ptr = aSig1<<shiftCount; - *zSig1Ptr = 0; - } - *zExpPtr = - shiftCount - 63; - } - else { - shiftCount = countLeadingZeros64( aSig0 ) - 15; - shortShift128Left( aSig0, aSig1, shiftCount, zSig0Ptr, zSig1Ptr ); - *zExpPtr = 1 - shiftCount; - } + int8 shiftCount; + + if ( aSig0 == 0 ) { + shiftCount = countLeadingZeros64( aSig1 ) - 15; + if ( shiftCount < 0 ) { + *zSig0Ptr = aSig1>>( - shiftCount ); + *zSig1Ptr = aSig1<<( shiftCount & 63 ); + } + else { + *zSig0Ptr = aSig1<<shiftCount; + *zSig1Ptr = 0; + } + *zExpPtr = - shiftCount - 63; + } + else { + shiftCount = countLeadingZeros64( aSig0 ) - 15; + shortShift128Left( aSig0, aSig1, shiftCount, zSig0Ptr, zSig1Ptr ); + *zExpPtr = 1 - shiftCount; + } } @@ -816,12 +816,12 @@ static void float32 int32_to_float32( int32 a ) { - flag zSign; + flag zSign; - if ( a == 0 ) return 0; - if ( a == (sbits32) 0x80000000 ) return packFloat32( 1, 0x9E, 0 ); - zSign = ( a < 0 ); - return normalizeRoundAndPackFloat32( zSign, 0x9C, zSign ? - a : a ); + if ( a == 0 ) return 0; + if ( a == (sbits32) 0x80000000 ) return packFloat32( 1, 0x9E, 0 ); + zSign = ( a < 0 ); + return normalizeRoundAndPackFloat32( zSign, 0x9C, zSign ? - a : a ); } @@ -833,17 +833,17 @@ float32 int32_to_float32( int32 a ) float64 int32_to_float64( int32 a ) { - flag zSign; - uint32 absA; - int8 shiftCount; - bits64 zSig; - - if ( a == 0 ) return 0; - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 21; - zSig = absA; - return packFloat64( zSign, 0x432 - shiftCount, zSig<<shiftCount ); + flag zSign; + uint32 absA; + int8 shiftCount; + bits64 zSig; + + if ( a == 0 ) return 0; + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros32( absA ) + 21; + zSig = absA; + return packFloat64( zSign, 0x432 - shiftCount, zSig<<shiftCount ); } @@ -858,17 +858,17 @@ float64 int32_to_float64( int32 a ) floatx80 int32_to_floatx80( int32 a ) { - flag zSign; - uint32 absA; - int8 shiftCount; - bits64 zSig; - - if ( a == 0 ) return packFloatx80( 0, 0, 0 ); - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 32; - zSig = absA; - return packFloatx80( zSign, 0x403E - shiftCount, zSig<<shiftCount ); + flag zSign; + uint32 absA; + int8 shiftCount; + bits64 zSig; + + if ( a == 0 ) return packFloatx80( 0, 0, 0 ); + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros32( absA ) + 32; + zSig = absA; + return packFloatx80( zSign, 0x403E - shiftCount, zSig<<shiftCount ); } @@ -884,17 +884,17 @@ floatx80 int32_to_floatx80( int32 a ) float128 int32_to_float128( int32 a ) { - flag zSign; - uint32 absA; - int8 shiftCount; - bits64 zSig0; - - if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros32( absA ) + 17; - zSig0 = absA; - return packFloat128( zSign, 0x402E - shiftCount, zSig0<<shiftCount, 0 ); + flag zSign; + uint32 absA; + int8 shiftCount; + bits64 zSig0; + + if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros32( absA ) + 17; + zSig0 = absA; + return packFloat128( zSign, 0x402E - shiftCount, zSig0<<shiftCount, 0 ); } @@ -908,28 +908,28 @@ float128 int32_to_float128( int32 a ) float32 int64_to_float32( int64 a ) { - flag zSign; - uint64 absA; - int8 shiftCount; + flag zSign; + uint64 absA; + int8 shiftCount; // bits32 zSig; - if ( a == 0 ) return 0; - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros64( absA ) - 40; - if ( 0 <= shiftCount ) { - return packFloat32( zSign, 0x95 - shiftCount, absA<<shiftCount ); - } - else { - shiftCount += 7; - if ( shiftCount < 0 ) { - shift64RightJamming( absA, - shiftCount, &absA ); - } - else { - absA <<= shiftCount; - } - return roundAndPackFloat32( zSign, 0x9C - shiftCount, absA ); - } + if ( a == 0 ) return 0; + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros64( absA ) - 40; + if ( 0 <= shiftCount ) { + return packFloat32( zSign, 0x95 - shiftCount, absA<<shiftCount ); + } + else { + shiftCount += 7; + if ( shiftCount < 0 ) { + shift64RightJamming( absA, - shiftCount, &absA ); + } + else { + absA <<= shiftCount; + } + return roundAndPackFloat32( zSign, 0x9C - shiftCount, absA ); + } } @@ -941,14 +941,14 @@ float32 int64_to_float32( int64 a ) float64 int64_to_float64( int64 a ) { - flag zSign; + flag zSign; - if ( a == 0 ) return 0; - if ( a == (sbits64) LIT64( 0x8000000000000000 ) ) { - return packFloat64( 1, 0x43E, 0 ); - } - zSign = ( a < 0 ); - return normalizeRoundAndPackFloat64( zSign, 0x43C, zSign ? - a : a ); + if ( a == 0 ) return 0; + if ( a == (sbits64) LIT64( 0x8000000000000000 ) ) { + return packFloat64( 1, 0x43E, 0 ); + } + zSign = ( a < 0 ); + return normalizeRoundAndPackFloat64( zSign, 0x43C, zSign ? - a : a ); } @@ -963,15 +963,15 @@ float64 int64_to_float64( int64 a ) floatx80 int64_to_floatx80( int64 a ) { - flag zSign; - uint64 absA; - int8 shiftCount; + flag zSign; + uint64 absA; + int8 shiftCount; - if ( a == 0 ) return packFloatx80( 0, 0, 0 ); - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros64( absA ); - return packFloatx80( zSign, 0x403E - shiftCount, absA<<shiftCount ); + if ( a == 0 ) return packFloatx80( 0, 0, 0 ); + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros64( absA ); + return packFloatx80( zSign, 0x403E - shiftCount, absA<<shiftCount ); } @@ -987,28 +987,28 @@ floatx80 int64_to_floatx80( int64 a ) float128 int64_to_float128( int64 a ) { - flag zSign; - uint64 absA; - int8 shiftCount; - int32 zExp; - bits64 zSig0, zSig1; - - if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); - zSign = ( a < 0 ); - absA = zSign ? - a : a; - shiftCount = countLeadingZeros64( absA ) + 49; - zExp = 0x406E - shiftCount; - if ( 64 <= shiftCount ) { - zSig1 = 0; - zSig0 = absA; - shiftCount -= 64; - } - else { - zSig1 = absA; - zSig0 = 0; - } - shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); - return packFloat128( zSign, zExp, zSig0, zSig1 ); + flag zSign; + uint64 absA; + int8 shiftCount; + int32 zExp; + bits64 zSig0, zSig1; + + if ( a == 0 ) return packFloat128( 0, 0, 0, 0 ); + zSign = ( a < 0 ); + absA = zSign ? - a : a; + shiftCount = countLeadingZeros64( absA ) + 49; + zExp = 0x406E - shiftCount; + if ( 64 <= shiftCount ) { + zSig1 = 0; + zSig0 = absA; + shiftCount -= 64; + } + else { + zSig1 = absA; + zSig0 = 0; + } + shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); + return packFloat128( zSign, zExp, zSig0, zSig1 ); } @@ -1026,21 +1026,21 @@ float128 int64_to_float128( int64 a ) int32 float32_to_int32( float32 a ) { - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - bits64 aSig64; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( ( aExp == 0xFF ) && aSig ) aSign = 0; - if ( aExp ) aSig |= 0x00800000; - shiftCount = 0xAF - aExp; - aSig64 = aSig; - aSig64 <<= 32; - if ( 0 < shiftCount ) shift64RightJamming( aSig64, shiftCount, &aSig64 ); - return roundAndPackInt32( aSign, aSig64 ); + flag aSign; + int16 aExp, shiftCount; + bits32 aSig; + bits64 aSig64; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + if ( ( aExp == 0xFF ) && aSig ) aSign = 0; + if ( aExp ) aSig |= 0x00800000; + shiftCount = 0xAF - aExp; + aSig64 = aSig; + aSig64 <<= 32; + if ( 0 < shiftCount ) shift64RightJamming( aSig64, shiftCount, &aSig64 ); + return roundAndPackInt32( aSign, aSig64 ); } @@ -1056,33 +1056,33 @@ int32 float32_to_int32( float32 a ) int32 float32_to_int32_round_to_zero( float32 a ) { - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - int32 z; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - shiftCount = aExp - 0x9E; - if ( 0 <= shiftCount ) { - if ( a != 0xCF000000 ) { - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) return 0x7FFFFFFF; - } - return (sbits32) 0x80000000; - } - else if ( aExp <= 0x7E ) { - if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - aSig = ( aSig | 0x00800000 )<<8; - z = aSig>>( - shiftCount ); - if ( (bits32) ( aSig<<( shiftCount & 31 ) ) ) { - float_exception_flags |= float_flag_inexact; - } - if ( aSign ) z = - z; - return z; + flag aSign; + int16 aExp, shiftCount; + bits32 aSig; + int32 z; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + shiftCount = aExp - 0x9E; + if ( 0 <= shiftCount ) { + if ( a != 0xCF000000 ) { + float_raise( float_flag_invalid ); + if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) return 0x7FFFFFFF; + } + return (sbits32) 0x80000000; + } + else if ( aExp <= 0x7E ) { + if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + aSig = ( aSig | 0x00800000 )<<8; + z = aSig>>( - shiftCount ); + if ( (bits32) ( aSig<<( shiftCount & 31 ) ) ) { + float_exception_flags |= float_flag_inexact; + } + if ( aSign ) z = - z; + return z; } @@ -1098,27 +1098,27 @@ int32 float32_to_int32_round_to_zero( float32 a ) int64 float32_to_int64( float32 a ) { - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - bits64 aSig64, aSigExtra; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - shiftCount = 0xBE - aExp; - if ( shiftCount < 0 ) { - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - if ( aExp ) aSig |= 0x00800000; - aSig64 = aSig; - aSig64 <<= 40; - shift64ExtraRightJamming( aSig64, 0, shiftCount, &aSig64, &aSigExtra ); - return roundAndPackInt64( aSign, aSig64, aSigExtra ); + flag aSign; + int16 aExp, shiftCount; + bits32 aSig; + bits64 aSig64, aSigExtra; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + shiftCount = 0xBE - aExp; + if ( shiftCount < 0 ) { + float_raise( float_flag_invalid ); + if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + if ( aExp ) aSig |= 0x00800000; + aSig64 = aSig; + aSig64 <<= 40; + shift64ExtraRightJamming( aSig64, 0, shiftCount, &aSig64, &aSigExtra ); + return roundAndPackInt64( aSign, aSig64, aSigExtra ); } @@ -1134,37 +1134,37 @@ int64 float32_to_int64( float32 a ) int64 float32_to_int64_round_to_zero( float32 a ) { - flag aSign; - int16 aExp, shiftCount; - bits32 aSig; - bits64 aSig64; - int64 z; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - shiftCount = aExp - 0xBE; - if ( 0 <= shiftCount ) { - if ( a != 0xDF000000 ) { - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - else if ( aExp <= 0x7E ) { - if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - aSig64 = aSig | 0x00800000; - aSig64 <<= 40; - z = aSig64>>( - shiftCount ); - if ( (bits64) ( aSig64<<( shiftCount & 63 ) ) ) { - float_exception_flags |= float_flag_inexact; - } - if ( aSign ) z = - z; - return z; + flag aSign; + int16 aExp, shiftCount; + bits32 aSig; + bits64 aSig64; + int64 z; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + shiftCount = aExp - 0xBE; + if ( 0 <= shiftCount ) { + if ( a != 0xDF000000 ) { + float_raise( float_flag_invalid ); + if ( ! aSign || ( ( aExp == 0xFF ) && aSig ) ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + else if ( aExp <= 0x7E ) { + if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + aSig64 = aSig | 0x00800000; + aSig64 <<= 40; + z = aSig64>>( - shiftCount ); + if ( (bits64) ( aSig64<<( shiftCount & 63 ) ) ) { + float_exception_flags |= float_flag_inexact; + } + if ( aSign ) z = - z; + return z; } @@ -1177,23 +1177,23 @@ int64 float32_to_int64_round_to_zero( float32 a ) float64 float32_to_float64( float32 a ) { - flag aSign; - int16 aExp; - bits32 aSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return commonNaNToFloat64( float32ToCommonNaN( a ) ); - return packFloat64( aSign, 0x7FF, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat64( aSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - --aExp; - } - return packFloat64( aSign, aExp + 0x380, ( (bits64) aSig )<<29 ); + flag aSign; + int16 aExp; + bits32 aSig; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + if ( aExp == 0xFF ) { + if ( aSig ) return commonNaNToFloat64( float32ToCommonNaN( a ) ); + return packFloat64( aSign, 0x7FF, 0 ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat64( aSign, 0, 0 ); + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + --aExp; + } + return packFloat64( aSign, aExp + 0x380, ( (bits64) aSig )<<29 ); } @@ -1208,23 +1208,23 @@ float64 float32_to_float64( float32 a ) floatx80 float32_to_floatx80( float32 a ) { - flag aSign; - int16 aExp; - bits32 aSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return commonNaNToFloatx80( float32ToCommonNaN( a ) ); - return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - aSig |= 0x00800000; - return packFloatx80( aSign, aExp + 0x3F80, ( (bits64) aSig )<<40 ); + flag aSign; + int16 aExp; + bits32 aSig; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + if ( aExp == 0xFF ) { + if ( aSig ) return commonNaNToFloatx80( float32ToCommonNaN( a ) ); + return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + } + aSig |= 0x00800000; + return packFloatx80( aSign, aExp + 0x3F80, ( (bits64) aSig )<<40 ); } @@ -1241,23 +1241,23 @@ floatx80 float32_to_floatx80( float32 a ) float128 float32_to_float128( float32 a ) { - flag aSign; - int16 aExp; - bits32 aSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return commonNaNToFloat128( float32ToCommonNaN( a ) ); - return packFloat128( aSign, 0x7FFF, 0, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - --aExp; - } - return packFloat128( aSign, aExp + 0x3F80, ( (bits64) aSig )<<25, 0 ); + flag aSign; + int16 aExp; + bits32 aSig; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + if ( aExp == 0xFF ) { + if ( aSig ) return commonNaNToFloat128( float32ToCommonNaN( a ) ); + return packFloat128( aSign, 0x7FFF, 0, 0 ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + --aExp; + } + return packFloat128( aSign, aExp + 0x3F80, ( (bits64) aSig )<<25, 0 ); } @@ -1272,53 +1272,53 @@ float128 float32_to_float128( float32 a ) float32 float32_round_to_int( float32 a ) { - flag aSign; - int16 aExp; - bits32 lastBitMask, roundBitsMask; - int8 roundingMode; - float32 z; - - aExp = extractFloat32Exp( a ); - if ( 0x96 <= aExp ) { - if ( ( aExp == 0xFF ) && extractFloat32Frac( a ) ) { - return propagateFloat32NaN( a, a ); - } - return a; - } - if ( aExp <= 0x7E ) { - if ( (bits32) ( a<<1 ) == 0 ) return a; - float_exception_flags |= float_flag_inexact; - aSign = extractFloat32Sign( a ); - switch ( float_rounding_mode ) { - case float_round_nearest_even: - if ( ( aExp == 0x7E ) && extractFloat32Frac( a ) ) { - return packFloat32( aSign, 0x7F, 0 ); - } - break; - case float_round_down: - return aSign ? 0xBF800000 : 0; - case float_round_up: - return aSign ? 0x80000000 : 0x3F800000; - } - return packFloat32( aSign, 0, 0 ); - } - lastBitMask = 1; - lastBitMask <<= 0x96 - aExp; - roundBitsMask = lastBitMask - 1; - z = a; - roundingMode = float_rounding_mode; - if ( roundingMode == float_round_nearest_even ) { - z += lastBitMask>>1; - if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; - } - else if ( roundingMode != float_round_to_zero ) { - if ( extractFloat32Sign( z ) ^ ( roundingMode == float_round_up ) ) { - z += roundBitsMask; - } - } - z &= ~ roundBitsMask; - if ( z != a ) float_exception_flags |= float_flag_inexact; - return z; + flag aSign; + int16 aExp; + bits32 lastBitMask, roundBitsMask; + int8 roundingMode; + float32 z; + + aExp = extractFloat32Exp( a ); + if ( 0x96 <= aExp ) { + if ( ( aExp == 0xFF ) && extractFloat32Frac( a ) ) { + return propagateFloat32NaN( a, a ); + } + return a; + } + if ( aExp <= 0x7E ) { + if ( (bits32) ( a<<1 ) == 0 ) return a; + float_exception_flags |= float_flag_inexact; + aSign = extractFloat32Sign( a ); + switch ( float_rounding_mode ) { + case float_round_nearest_even: + if ( ( aExp == 0x7E ) && extractFloat32Frac( a ) ) { + return packFloat32( aSign, 0x7F, 0 ); + } + break; + case float_round_down: + return aSign ? 0xBF800000 : 0; + case float_round_up: + return aSign ? 0x80000000 : 0x3F800000; + } + return packFloat32( aSign, 0, 0 ); + } + lastBitMask = 1; + lastBitMask <<= 0x96 - aExp; + roundBitsMask = lastBitMask - 1; + z = a; + roundingMode = float_rounding_mode; + if ( roundingMode == float_round_nearest_even ) { + z += lastBitMask>>1; + if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; + } + else if ( roundingMode != float_round_to_zero ) { + if ( extractFloat32Sign( z ) ^ ( roundingMode == float_round_up ) ) { + z += roundBitsMask; + } + } + z &= ~ roundBitsMask; + if ( z != a ) float_exception_flags |= float_flag_inexact; + return z; } @@ -1332,64 +1332,64 @@ float32 float32_round_to_int( float32 a ) static float32 addFloat32Sigs( float32 a, float32 b, flag zSign ) { - int16 aExp, bExp, zExp; - bits32 aSig, bSig, zSig; - int16 expDiff; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - bSig = extractFloat32Frac( b ); - bExp = extractFloat32Exp( b ); - expDiff = aExp - bExp; - aSig <<= 6; - bSig <<= 6; - if ( 0 < expDiff ) { - if ( aExp == 0xFF ) { - if ( aSig ) return propagateFloat32NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig |= 0x20000000; - } - shift32RightJamming( bSig, expDiff, &bSig ); - zExp = aExp; - } - else if ( expDiff < 0 ) { - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - return packFloat32( zSign, 0xFF, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig |= 0x20000000; - } - shift32RightJamming( aSig, - expDiff, &aSig ); - zExp = bExp; - } - else { - if ( aExp == 0xFF ) { - if ( aSig | bSig ) return propagateFloat32NaN( a, b ); - return a; - } - if ( aExp == 0 ) return packFloat32( zSign, 0, ( aSig + bSig )>>6 ); - zSig = 0x40000000 + aSig + bSig; - zExp = aExp; - goto roundAndPack; - } - aSig |= 0x20000000; - zSig = ( aSig + bSig )<<1; - --zExp; - if ( (sbits32) zSig < 0 ) { - zSig = aSig + bSig; - ++zExp; - } - roundAndPack: - return roundAndPackFloat32( zSign, zExp, zSig ); + int16 aExp, bExp, zExp; + bits32 aSig, bSig, zSig; + int16 expDiff; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + bSig = extractFloat32Frac( b ); + bExp = extractFloat32Exp( b ); + expDiff = aExp - bExp; + aSig <<= 6; + bSig <<= 6; + if ( 0 < expDiff ) { + if ( aExp == 0xFF ) { + if ( aSig ) return propagateFloat32NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig |= 0x20000000; + } + shift32RightJamming( bSig, expDiff, &bSig ); + zExp = aExp; + } + else if ( expDiff < 0 ) { + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + return packFloat32( zSign, 0xFF, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig |= 0x20000000; + } + shift32RightJamming( aSig, - expDiff, &aSig ); + zExp = bExp; + } + else { + if ( aExp == 0xFF ) { + if ( aSig | bSig ) return propagateFloat32NaN( a, b ); + return a; + } + if ( aExp == 0 ) return packFloat32( zSign, 0, ( aSig + bSig )>>6 ); + zSig = 0x40000000 + aSig + bSig; + zExp = aExp; + goto roundAndPack; + } + aSig |= 0x20000000; + zSig = ( aSig + bSig )<<1; + --zExp; + if ( (sbits32) zSig < 0 ) { + zSig = aSig + bSig; + ++zExp; + } + roundAndPack: + return roundAndPackFloat32( zSign, zExp, zSig ); } @@ -1403,68 +1403,68 @@ static float32 addFloat32Sigs( float32 a, float32 b, flag zSign ) static float32 subFloat32Sigs( float32 a, float32 b, flag zSign ) { - int16 aExp, bExp, zExp; - bits32 aSig, bSig, zSig; - int16 expDiff; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - bSig = extractFloat32Frac( b ); - bExp = extractFloat32Exp( b ); - expDiff = aExp - bExp; - aSig <<= 7; - bSig <<= 7; - if ( 0 < expDiff ) goto aExpBigger; - if ( expDiff < 0 ) goto bExpBigger; - if ( aExp == 0xFF ) { - if ( aSig | bSig ) return propagateFloat32NaN( a, b ); - float_raise( float_flag_invalid ); - return float32_default_nan; - } - if ( aExp == 0 ) { - aExp = 1; - bExp = 1; - } - if ( bSig < aSig ) goto aBigger; - if ( aSig < bSig ) goto bBigger; - return packFloat32( float_rounding_mode == float_round_down, 0, 0 ); - bExpBigger: - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - return packFloat32( zSign ^ 1, 0xFF, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig |= 0x40000000; - } - shift32RightJamming( aSig, - expDiff, &aSig ); - bSig |= 0x40000000; - bBigger: - zSig = bSig - aSig; - zExp = bExp; - zSign ^= 1; - goto normalizeRoundAndPack; - aExpBigger: - if ( aExp == 0xFF ) { - if ( aSig ) return propagateFloat32NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig |= 0x40000000; - } - shift32RightJamming( bSig, expDiff, &bSig ); - aSig |= 0x40000000; - aBigger: - zSig = aSig - bSig; - zExp = aExp; - normalizeRoundAndPack: - --zExp; - return normalizeRoundAndPackFloat32( zSign, zExp, zSig ); + int16 aExp, bExp, zExp; + bits32 aSig, bSig, zSig; + int16 expDiff; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + bSig = extractFloat32Frac( b ); + bExp = extractFloat32Exp( b ); + expDiff = aExp - bExp; + aSig <<= 7; + bSig <<= 7; + if ( 0 < expDiff ) goto aExpBigger; + if ( expDiff < 0 ) goto bExpBigger; + if ( aExp == 0xFF ) { + if ( aSig | bSig ) return propagateFloat32NaN( a, b ); + float_raise( float_flag_invalid ); + return float32_default_nan; + } + if ( aExp == 0 ) { + aExp = 1; + bExp = 1; + } + if ( bSig < aSig ) goto aBigger; + if ( aSig < bSig ) goto bBigger; + return packFloat32( float_rounding_mode == float_round_down, 0, 0 ); + bExpBigger: + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + return packFloat32( zSign ^ 1, 0xFF, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig |= 0x40000000; + } + shift32RightJamming( aSig, - expDiff, &aSig ); + bSig |= 0x40000000; + bBigger: + zSig = bSig - aSig; + zExp = bExp; + zSign ^= 1; + goto normalizeRoundAndPack; + aExpBigger: + if ( aExp == 0xFF ) { + if ( aSig ) return propagateFloat32NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig |= 0x40000000; + } + shift32RightJamming( bSig, expDiff, &bSig ); + aSig |= 0x40000000; + aBigger: + zSig = aSig - bSig; + zExp = aExp; + normalizeRoundAndPack: + --zExp; + return normalizeRoundAndPackFloat32( zSign, zExp, zSig ); } @@ -1476,16 +1476,16 @@ static float32 subFloat32Sigs( float32 a, float32 b, flag zSign ) float32 float32_add( float32 a, float32 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign == bSign ) { - return addFloat32Sigs( a, b, aSign ); - } - else { - return subFloat32Sigs( a, b, aSign ); - } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign == bSign ) { + return addFloat32Sigs( a, b, aSign ); + } + else { + return subFloat32Sigs( a, b, aSign ); + } } @@ -1497,16 +1497,16 @@ float32 float32_add( float32 a, float32 b ) float32 float32_sub( float32 a, float32 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign == bSign ) { - return subFloat32Sigs( a, b, aSign ); - } - else { - return addFloat32Sigs( a, b, aSign ); - } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign == bSign ) { + return subFloat32Sigs( a, b, aSign ); + } + else { + return addFloat32Sigs( a, b, aSign ); + } } @@ -1518,55 +1518,55 @@ float32 float32_sub( float32 a, float32 b ) float32 float32_mul( float32 a, float32 b ) { - flag aSign, bSign, zSign; - int16 aExp, bExp, zExp; - bits32 aSig, bSig; - bits64 zSig64; - bits32 zSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - bSig = extractFloat32Frac( b ); - bExp = extractFloat32Exp( b ); - bSign = extractFloat32Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0xFF ) { - if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { - return propagateFloat32NaN( a, b ); - } - if ( ( bExp | bSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float32_default_nan; - } - return packFloat32( zSign, 0xFF, 0 ); - } - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - if ( ( aExp | aSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float32_default_nan; - } - return packFloat32( zSign, 0xFF, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) return packFloat32( zSign, 0, 0 ); - normalizeFloat32Subnormal( bSig, &bExp, &bSig ); - } - zExp = aExp + bExp - 0x7F; - aSig = ( aSig | 0x00800000 )<<7; - bSig = ( bSig | 0x00800000 )<<8; - shift64RightJamming( ( (bits64) aSig ) * bSig, 32, &zSig64 ); - zSig = zSig64; - if ( 0 <= (sbits32) ( zSig<<1 ) ) { - zSig <<= 1; - --zExp; - } - return roundAndPackFloat32( zSign, zExp, zSig ); + flag aSign, bSign, zSign; + int16 aExp, bExp, zExp; + bits32 aSig, bSig; + bits64 zSig64; + bits32 zSig; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + bSig = extractFloat32Frac( b ); + bExp = extractFloat32Exp( b ); + bSign = extractFloat32Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0xFF ) { + if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { + return propagateFloat32NaN( a, b ); + } + if ( ( bExp | bSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float32_default_nan; + } + return packFloat32( zSign, 0xFF, 0 ); + } + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + if ( ( aExp | aSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float32_default_nan; + } + return packFloat32( zSign, 0xFF, 0 ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) return packFloat32( zSign, 0, 0 ); + normalizeFloat32Subnormal( bSig, &bExp, &bSig ); + } + zExp = aExp + bExp - 0x7F; + aSig = ( aSig | 0x00800000 )<<7; + bSig = ( bSig | 0x00800000 )<<8; + shift64RightJamming( ( (bits64) aSig ) * bSig, 32, &zSig64 ); + zSig = zSig64; + if ( 0 <= (sbits32) ( zSig<<1 ) ) { + zSig <<= 1; + --zExp; + } + return roundAndPackFloat32( zSign, zExp, zSig ); } @@ -1578,57 +1578,57 @@ float32 float32_mul( float32 a, float32 b ) float32 float32_div( float32 a, float32 b ) { - flag aSign, bSign, zSign; - int16 aExp, bExp, zExp; - bits32 aSig, bSig, zSig; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - bSig = extractFloat32Frac( b ); - bExp = extractFloat32Exp( b ); - bSign = extractFloat32Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0xFF ) { - if ( aSig ) return propagateFloat32NaN( a, b ); - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - float_raise( float_flag_invalid ); - return float32_default_nan; - } - return packFloat32( zSign, 0xFF, 0 ); - } - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - return packFloat32( zSign, 0, 0 ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - if ( ( aExp | aSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float32_default_nan; - } - float_raise( float_flag_divbyzero ); - return packFloat32( zSign, 0xFF, 0 ); - } - normalizeFloat32Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - zExp = aExp - bExp + 0x7D; - aSig = ( aSig | 0x00800000 )<<7; - bSig = ( bSig | 0x00800000 )<<8; - if ( bSig <= ( aSig + aSig ) ) { - aSig >>= 1; - ++zExp; - } - zSig = ( ( (bits64) aSig )<<32 ) / bSig; - if ( ( zSig & 0x3F ) == 0 ) { - zSig |= ( (bits64) bSig * zSig != ( (bits64) aSig )<<32 ); - } - return roundAndPackFloat32( zSign, zExp, zSig ); + flag aSign, bSign, zSign; + int16 aExp, bExp, zExp; + bits32 aSig, bSig, zSig; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + bSig = extractFloat32Frac( b ); + bExp = extractFloat32Exp( b ); + bSign = extractFloat32Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0xFF ) { + if ( aSig ) return propagateFloat32NaN( a, b ); + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + float_raise( float_flag_invalid ); + return float32_default_nan; + } + return packFloat32( zSign, 0xFF, 0 ); + } + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + return packFloat32( zSign, 0, 0 ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + if ( ( aExp | aSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float32_default_nan; + } + float_raise( float_flag_divbyzero ); + return packFloat32( zSign, 0xFF, 0 ); + } + normalizeFloat32Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat32( zSign, 0, 0 ); + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + } + zExp = aExp - bExp + 0x7D; + aSig = ( aSig | 0x00800000 )<<7; + bSig = ( bSig | 0x00800000 )<<8; + if ( bSig <= ( aSig + aSig ) ) { + aSig >>= 1; + ++zExp; + } + zSig = ( ( (bits64) aSig )<<32 ) / bSig; + if ( ( zSig & 0x3F ) == 0 ) { + zSig |= ( (bits64) bSig * zSig != ( (bits64) aSig )<<32 ); + } + return roundAndPackFloat32( zSign, zExp, zSig ); } @@ -1640,95 +1640,95 @@ float32 float32_div( float32 a, float32 b ) float32 float32_rem( float32 a, float32 b ) { - flag aSign, zSign; - int16 aExp, bExp, expDiff; - bits32 aSig, bSig; - bits32 q; - bits64 aSig64, bSig64, q64; - bits32 alternateASig; - sbits32 sigMean; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - bSig = extractFloat32Frac( b ); - bExp = extractFloat32Exp( b ); + flag aSign, zSign; + int16 aExp, bExp, expDiff; + bits32 aSig, bSig; + bits32 q; + bits64 aSig64, bSig64, q64; + bits32 alternateASig; + sbits32 sigMean; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + bSig = extractFloat32Frac( b ); + bExp = extractFloat32Exp( b ); // bSign = extractFloat32Sign( b ); - if ( aExp == 0xFF ) { - if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { - return propagateFloat32NaN( a, b ); - } - float_raise( float_flag_invalid ); - return float32_default_nan; - } - if ( bExp == 0xFF ) { - if ( bSig ) return propagateFloat32NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - float_raise( float_flag_invalid ); - return float32_default_nan; - } - normalizeFloat32Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return a; - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - expDiff = aExp - bExp; - aSig |= 0x00800000; - bSig |= 0x00800000; - if ( expDiff < 32 ) { - aSig <<= 8; - bSig <<= 8; - if ( expDiff < 0 ) { - if ( expDiff < -1 ) return a; - aSig >>= 1; - } - q = ( bSig <= aSig ); - if ( q ) aSig -= bSig; - if ( 0 < expDiff ) { - q = ( ( (bits64) aSig )<<32 ) / bSig; - q >>= 32 - expDiff; - bSig >>= 2; - aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; - } - else { - aSig >>= 2; - bSig >>= 2; - } - } - else { - if ( bSig <= aSig ) aSig -= bSig; - aSig64 = ( (bits64) aSig )<<40; - bSig64 = ( (bits64) bSig )<<40; - expDiff -= 64; - while ( 0 < expDiff ) { - q64 = estimateDiv128To64( aSig64, 0, bSig64 ); - q64 = ( 2 < q64 ) ? q64 - 2 : 0; - aSig64 = - ( ( bSig * q64 )<<38 ); - expDiff -= 62; - } - expDiff += 64; - q64 = estimateDiv128To64( aSig64, 0, bSig64 ); - q64 = ( 2 < q64 ) ? q64 - 2 : 0; - q = q64>>( 64 - expDiff ); - bSig <<= 6; - aSig = ( ( aSig64>>33 )<<( expDiff - 1 ) ) - bSig * q; - } - do { - alternateASig = aSig; - ++q; - aSig -= bSig; - } while ( 0 <= (sbits32) aSig ); - sigMean = aSig + alternateASig; - if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { - aSig = alternateASig; - } - zSign = ( (sbits32) aSig < 0 ); - if ( zSign ) aSig = - aSig; - return normalizeRoundAndPackFloat32( aSign ^ zSign, bExp, aSig ); + if ( aExp == 0xFF ) { + if ( aSig || ( ( bExp == 0xFF ) && bSig ) ) { + return propagateFloat32NaN( a, b ); + } + float_raise( float_flag_invalid ); + return float32_default_nan; + } + if ( bExp == 0xFF ) { + if ( bSig ) return propagateFloat32NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + float_raise( float_flag_invalid ); + return float32_default_nan; + } + normalizeFloat32Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return a; + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + } + expDiff = aExp - bExp; + aSig |= 0x00800000; + bSig |= 0x00800000; + if ( expDiff < 32 ) { + aSig <<= 8; + bSig <<= 8; + if ( expDiff < 0 ) { + if ( expDiff < -1 ) return a; + aSig >>= 1; + } + q = ( bSig <= aSig ); + if ( q ) aSig -= bSig; + if ( 0 < expDiff ) { + q = ( ( (bits64) aSig )<<32 ) / bSig; + q >>= 32 - expDiff; + bSig >>= 2; + aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; + } + else { + aSig >>= 2; + bSig >>= 2; + } + } + else { + if ( bSig <= aSig ) aSig -= bSig; + aSig64 = ( (bits64) aSig )<<40; + bSig64 = ( (bits64) bSig )<<40; + expDiff -= 64; + while ( 0 < expDiff ) { + q64 = estimateDiv128To64( aSig64, 0, bSig64 ); + q64 = ( 2 < q64 ) ? q64 - 2 : 0; + aSig64 = - ( ( bSig * q64 )<<38 ); + expDiff -= 62; + } + expDiff += 64; + q64 = estimateDiv128To64( aSig64, 0, bSig64 ); + q64 = ( 2 < q64 ) ? q64 - 2 : 0; + q = q64>>( 64 - expDiff ); + bSig <<= 6; + aSig = ( ( aSig64>>33 )<<( expDiff - 1 ) ) - bSig * q; + } + do { + alternateASig = aSig; + ++q; + aSig -= bSig; + } while ( 0 <= (sbits32) aSig ); + sigMean = aSig + alternateASig; + if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { + aSig = alternateASig; + } + zSign = ( (sbits32) aSig < 0 ); + if ( zSign ) aSig = - aSig; + return normalizeRoundAndPackFloat32( aSign ^ zSign, bExp, aSig ); } @@ -1740,49 +1740,49 @@ float32 float32_rem( float32 a, float32 b ) float32 float32_sqrt( float32 a ) { - flag aSign; - int16 aExp, zExp; - bits32 aSig, zSig; - bits64 rem, term; - - aSig = extractFloat32Frac( a ); - aExp = extractFloat32Exp( a ); - aSign = extractFloat32Sign( a ); - if ( aExp == 0xFF ) { - if ( aSig ) return propagateFloat32NaN( a, 0 ); - if ( ! aSign ) return a; - float_raise( float_flag_invalid ); - return float32_default_nan; - } - if ( aSign ) { - if ( ( aExp | aSig ) == 0 ) return a; - float_raise( float_flag_invalid ); - return float32_default_nan; - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return 0; - normalizeFloat32Subnormal( aSig, &aExp, &aSig ); - } - zExp = ( ( aExp - 0x7F )>>1 ) + 0x7E; - aSig = ( aSig | 0x00800000 )<<8; - zSig = estimateSqrt32( aExp, aSig ) + 2; - if ( ( zSig & 0x7F ) <= 5 ) { - if ( zSig < 2 ) { - zSig = 0x7FFFFFFF; - goto roundAndPack; - } - aSig >>= aExp & 1; - term = ( (bits64) zSig ) * zSig; - rem = ( ( (bits64) aSig )<<32 ) - term; - while ( (sbits64) rem < 0 ) { - --zSig; - rem += ( ( (bits64) zSig )<<1 ) | 1; - } - zSig |= ( rem != 0 ); - } - shift32RightJamming( zSig, 1, &zSig ); - roundAndPack: - return roundAndPackFloat32( 0, zExp, zSig ); + flag aSign; + int16 aExp, zExp; + bits32 aSig, zSig; + bits64 rem, term; + + aSig = extractFloat32Frac( a ); + aExp = extractFloat32Exp( a ); + aSign = extractFloat32Sign( a ); + if ( aExp == 0xFF ) { + if ( aSig ) return propagateFloat32NaN( a, 0 ); + if ( ! aSign ) return a; + float_raise( float_flag_invalid ); + return float32_default_nan; + } + if ( aSign ) { + if ( ( aExp | aSig ) == 0 ) return a; + float_raise( float_flag_invalid ); + return float32_default_nan; + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return 0; + normalizeFloat32Subnormal( aSig, &aExp, &aSig ); + } + zExp = ( ( aExp - 0x7F )>>1 ) + 0x7E; + aSig = ( aSig | 0x00800000 )<<8; + zSig = estimateSqrt32( aExp, aSig ) + 2; + if ( ( zSig & 0x7F ) <= 5 ) { + if ( zSig < 2 ) { + zSig = 0x7FFFFFFF; + goto roundAndPack; + } + aSig >>= aExp & 1; + term = ( (bits64) zSig ) * zSig; + rem = ( ( (bits64) aSig )<<32 ) - term; + while ( (sbits64) rem < 0 ) { + --zSig; + rem += ( ( (bits64) zSig )<<1 ) | 1; + } + zSig |= ( rem != 0 ); + } + shift32RightJamming( zSig, 1, &zSig ); + roundAndPack: + return roundAndPackFloat32( 0, zExp, zSig ); } @@ -1795,15 +1795,15 @@ float32 float32_sqrt( float32 a ) flag float32_eq( float32 a, float32 b ) { - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 ); + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 ); } @@ -1816,18 +1816,18 @@ flag float32_eq( float32 a, float32 b ) flag float32_le( float32 a, float32 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 ); - return ( a == b ) || ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 ); + return ( a == b ) || ( aSign ^ ( a < b ) ); } @@ -1839,18 +1839,18 @@ flag float32_le( float32 a, float32 b ) flag float32_lt( float32 a, float32 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 ); - return ( a != b ) && ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 ); + return ( a != b ) && ( aSign ^ ( a < b ) ); } @@ -1864,13 +1864,13 @@ flag float32_lt( float32 a, float32 b ) flag float32_eq_signaling( float32 a, float32 b ) { - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 ); + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + return ( a == b ) || ( (bits32) ( ( a | b )<<1 ) == 0 ); } @@ -1883,21 +1883,21 @@ flag float32_eq_signaling( float32 a, float32 b ) flag float32_le_quiet( float32 a, float32 b ) { - flag aSign, bSign; + flag aSign, bSign; // int16 aExp, bExp; - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 ); - return ( a == b ) || ( aSign ^ ( a < b ) ); + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign != bSign ) return aSign || ( (bits32) ( ( a | b )<<1 ) == 0 ); + return ( a == b ) || ( aSign ^ ( a < b ) ); } @@ -1910,20 +1910,20 @@ flag float32_le_quiet( float32 a, float32 b ) flag float32_lt_quiet( float32 a, float32 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) - || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) - ) { - if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat32Sign( a ); - bSign = extractFloat32Sign( b ); - if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 ); - return ( a != b ) && ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat32Exp( a ) == 0xFF ) && extractFloat32Frac( a ) ) + || ( ( extractFloat32Exp( b ) == 0xFF ) && extractFloat32Frac( b ) ) + ) { + if ( float32_is_signaling_nan( a ) || float32_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat32Sign( a ); + bSign = extractFloat32Sign( b ); + if ( aSign != bSign ) return aSign && ( (bits32) ( ( a | b )<<1 ) != 0 ); + return ( a != b ) && ( aSign ^ ( a < b ) ); } @@ -1939,18 +1939,18 @@ flag float32_lt_quiet( float32 a, float32 b ) int32 float64_to_int32( float64 a ) { - flag aSign; - int16 aExp, shiftCount; - bits64 aSig; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; - if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); - shiftCount = 0x42C - aExp; - if ( 0 < shiftCount ) shift64RightJamming( aSig, shiftCount, &aSig ); - return roundAndPackInt32( aSign, aSig ); + flag aSign; + int16 aExp, shiftCount; + bits64 aSig; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; + if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); + shiftCount = 0x42C - aExp; + if ( 0 < shiftCount ) shift64RightJamming( aSig, shiftCount, &aSig ); + return roundAndPackInt32( aSign, aSig ); } @@ -1966,37 +1966,37 @@ int32 float64_to_int32( float64 a ) int32 float64_to_int32_round_to_zero( float64 a ) { - flag aSign; - int16 aExp, shiftCount; - bits64 aSig, savedASig; - int32 z; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( 0x41E < aExp ) { - if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; - goto invalid; - } - else if ( aExp < 0x3FF ) { - if ( aExp || aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - aSig |= LIT64( 0x0010000000000000 ); - shiftCount = 0x433 - aExp; - savedASig = aSig; - aSig >>= shiftCount; - z = aSig; - if ( aSign ) z = - z; - if ( ( z < 0 ) ^ aSign ) { - invalid: - float_raise( float_flag_invalid ); - return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; - } - if ( ( aSig<<shiftCount ) != savedASig ) { - float_exception_flags |= float_flag_inexact; - } - return z; + flag aSign; + int16 aExp, shiftCount; + bits64 aSig, savedASig; + int32 z; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( 0x41E < aExp ) { + if ( ( aExp == 0x7FF ) && aSig ) aSign = 0; + goto invalid; + } + else if ( aExp < 0x3FF ) { + if ( aExp || aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + aSig |= LIT64( 0x0010000000000000 ); + shiftCount = 0x433 - aExp; + savedASig = aSig; + aSig >>= shiftCount; + z = aSig; + if ( aSign ) z = - z; + if ( ( z < 0 ) ^ aSign ) { + invalid: + float_raise( float_flag_invalid ); + return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; + } + if ( ( aSig<<shiftCount ) != savedASig ) { + float_exception_flags |= float_flag_inexact; + } + return z; } @@ -2012,33 +2012,33 @@ int32 float64_to_int32_round_to_zero( float64 a ) int64 float64_to_int64( float64 a ) { - flag aSign; - int16 aExp, shiftCount; - bits64 aSig, aSigExtra; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); - shiftCount = 0x433 - aExp; - if ( shiftCount <= 0 ) { - if ( 0x43E < aExp ) { - float_raise( float_flag_invalid ); - if ( ! aSign - || ( ( aExp == 0x7FF ) - && ( aSig != LIT64( 0x0010000000000000 ) ) ) - ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - aSigExtra = 0; - aSig <<= - shiftCount; - } - else { - shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra ); - } - return roundAndPackInt64( aSign, aSig, aSigExtra ); + flag aSign; + int16 aExp, shiftCount; + bits64 aSig, aSigExtra; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); + shiftCount = 0x433 - aExp; + if ( shiftCount <= 0 ) { + if ( 0x43E < aExp ) { + float_raise( float_flag_invalid ); + if ( ! aSign + || ( ( aExp == 0x7FF ) + && ( aSig != LIT64( 0x0010000000000000 ) ) ) + ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + aSigExtra = 0; + aSig <<= - shiftCount; + } + else { + shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra ); + } + return roundAndPackInt64( aSign, aSig, aSigExtra ); } @@ -2054,43 +2054,43 @@ int64 float64_to_int64( float64 a ) int64 float64_to_int64_round_to_zero( float64 a ) { - flag aSign; - int16 aExp, shiftCount; - bits64 aSig; - int64 z; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); - shiftCount = aExp - 0x433; - if ( 0 <= shiftCount ) { - if ( 0x43E <= aExp ) { - if ( a != LIT64( 0xC3E0000000000000 ) ) { - float_raise( float_flag_invalid ); - if ( ! aSign - || ( ( aExp == 0x7FF ) - && ( aSig != LIT64( 0x0010000000000000 ) ) ) - ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - z = aSig<<shiftCount; - } - else { - if ( aExp < 0x3FE ) { - if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - z = aSig>>( - shiftCount ); - if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { - float_exception_flags |= float_flag_inexact; - } - } - if ( aSign ) z = - z; - return z; + flag aSign; + int16 aExp, shiftCount; + bits64 aSig; + int64 z; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp ) aSig |= LIT64( 0x0010000000000000 ); + shiftCount = aExp - 0x433; + if ( 0 <= shiftCount ) { + if ( 0x43E <= aExp ) { + if ( a != LIT64( 0xC3E0000000000000 ) ) { + float_raise( float_flag_invalid ); + if ( ! aSign + || ( ( aExp == 0x7FF ) + && ( aSig != LIT64( 0x0010000000000000 ) ) ) + ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + z = aSig<<shiftCount; + } + else { + if ( aExp < 0x3FE ) { + if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + z = aSig>>( - shiftCount ); + if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { + float_exception_flags |= float_flag_inexact; + } + } + if ( aSign ) z = - z; + return z; } @@ -2103,25 +2103,25 @@ int64 float64_to_int64_round_to_zero( float64 a ) float32 float64_to_float32( float64 a ) { - flag aSign; - int16 aExp; - bits64 aSig; - bits32 zSig; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp == 0x7FF ) { - if ( aSig ) return commonNaNToFloat32( float64ToCommonNaN( a ) ); - return packFloat32( aSign, 0xFF, 0 ); - } - shift64RightJamming( aSig, 22, &aSig ); - zSig = aSig; - if ( aExp || zSig ) { - zSig |= 0x40000000; - aExp -= 0x381; - } - return roundAndPackFloat32( aSign, aExp, zSig ); + flag aSign; + int16 aExp; + bits64 aSig; + bits32 zSig; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp == 0x7FF ) { + if ( aSig ) return commonNaNToFloat32( float64ToCommonNaN( a ) ); + return packFloat32( aSign, 0xFF, 0 ); + } + shift64RightJamming( aSig, 22, &aSig ); + zSig = aSig; + if ( aExp || zSig ) { + zSig |= 0x40000000; + aExp -= 0x381; + } + return roundAndPackFloat32( aSign, aExp, zSig ); } @@ -2136,24 +2136,24 @@ float32 float64_to_float32( float64 a ) floatx80 float64_to_floatx80( float64 a ) { - flag aSign; - int16 aExp; - bits64 aSig; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp == 0x7FF ) { - if ( aSig ) return commonNaNToFloatx80( float64ToCommonNaN( a ) ); - return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - } - return - packFloatx80( - aSign, aExp + 0x3C00, ( aSig | LIT64( 0x0010000000000000 ) )<<11 ); + flag aSign; + int16 aExp; + bits64 aSig; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp == 0x7FF ) { + if ( aSig ) return commonNaNToFloatx80( float64ToCommonNaN( a ) ); + return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloatx80( aSign, 0, 0 ); + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + } + return + packFloatx80( + aSign, aExp + 0x3C00, ( aSig | LIT64( 0x0010000000000000 ) )<<11 ); } @@ -2170,24 +2170,24 @@ floatx80 float64_to_floatx80( float64 a ) float128 float64_to_float128( float64 a ) { - flag aSign; - int16 aExp; - bits64 aSig, zSig0, zSig1; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp == 0x7FF ) { - if ( aSig ) return commonNaNToFloat128( float64ToCommonNaN( a ) ); - return packFloat128( aSign, 0x7FFF, 0, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - --aExp; - } - shift128Right( aSig, 0, 4, &zSig0, &zSig1 ); - return packFloat128( aSign, aExp + 0x3C00, zSig0, zSig1 ); + flag aSign; + int16 aExp; + bits64 aSig, zSig0, zSig1; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp == 0x7FF ) { + if ( aSig ) return commonNaNToFloat128( float64ToCommonNaN( a ) ); + return packFloat128( aSign, 0x7FFF, 0, 0 ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat128( aSign, 0, 0, 0 ); + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + --aExp; + } + shift128Right( aSig, 0, 4, &zSig0, &zSig1 ); + return packFloat128( aSign, aExp + 0x3C00, zSig0, zSig1 ); } @@ -2202,54 +2202,54 @@ float128 float64_to_float128( float64 a ) float64 float64_round_to_int( float64 a ) { - flag aSign; - int16 aExp; - bits64 lastBitMask, roundBitsMask; - int8 roundingMode; - float64 z; - - aExp = extractFloat64Exp( a ); - if ( 0x433 <= aExp ) { - if ( ( aExp == 0x7FF ) && extractFloat64Frac( a ) ) { - return propagateFloat64NaN( a, a ); - } - return a; - } - if ( aExp < 0x3FF ) { - if ( (bits64) ( a<<1 ) == 0 ) return a; - float_exception_flags |= float_flag_inexact; - aSign = extractFloat64Sign( a ); - switch ( float_rounding_mode ) { - case float_round_nearest_even: - if ( ( aExp == 0x3FE ) && extractFloat64Frac( a ) ) { - return packFloat64( aSign, 0x3FF, 0 ); - } - break; - case float_round_down: - return aSign ? LIT64( 0xBFF0000000000000 ) : 0; - case float_round_up: - return - aSign ? LIT64( 0x8000000000000000 ) : LIT64( 0x3FF0000000000000 ); - } - return packFloat64( aSign, 0, 0 ); - } - lastBitMask = 1; - lastBitMask <<= 0x433 - aExp; - roundBitsMask = lastBitMask - 1; - z = a; - roundingMode = float_rounding_mode; - if ( roundingMode == float_round_nearest_even ) { - z += lastBitMask>>1; - if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; - } - else if ( roundingMode != float_round_to_zero ) { - if ( extractFloat64Sign( z ) ^ ( roundingMode == float_round_up ) ) { - z += roundBitsMask; - } - } - z &= ~ roundBitsMask; - if ( z != a ) float_exception_flags |= float_flag_inexact; - return z; + flag aSign; + int16 aExp; + bits64 lastBitMask, roundBitsMask; + int8 roundingMode; + float64 z; + + aExp = extractFloat64Exp( a ); + if ( 0x433 <= aExp ) { + if ( ( aExp == 0x7FF ) && extractFloat64Frac( a ) ) { + return propagateFloat64NaN( a, a ); + } + return a; + } + if ( aExp < 0x3FF ) { + if ( (bits64) ( a<<1 ) == 0 ) return a; + float_exception_flags |= float_flag_inexact; + aSign = extractFloat64Sign( a ); + switch ( float_rounding_mode ) { + case float_round_nearest_even: + if ( ( aExp == 0x3FE ) && extractFloat64Frac( a ) ) { + return packFloat64( aSign, 0x3FF, 0 ); + } + break; + case float_round_down: + return aSign ? LIT64( 0xBFF0000000000000 ) : 0; + case float_round_up: + return + aSign ? LIT64( 0x8000000000000000 ) : LIT64( 0x3FF0000000000000 ); + } + return packFloat64( aSign, 0, 0 ); + } + lastBitMask = 1; + lastBitMask <<= 0x433 - aExp; + roundBitsMask = lastBitMask - 1; + z = a; + roundingMode = float_rounding_mode; + if ( roundingMode == float_round_nearest_even ) { + z += lastBitMask>>1; + if ( ( z & roundBitsMask ) == 0 ) z &= ~ lastBitMask; + } + else if ( roundingMode != float_round_to_zero ) { + if ( extractFloat64Sign( z ) ^ ( roundingMode == float_round_up ) ) { + z += roundBitsMask; + } + } + z &= ~ roundBitsMask; + if ( z != a ) float_exception_flags |= float_flag_inexact; + return z; } @@ -2263,64 +2263,64 @@ float64 float64_round_to_int( float64 a ) static float64 addFloat64Sigs( float64 a, float64 b, flag zSign ) { - int16 aExp, bExp, zExp; - bits64 aSig, bSig, zSig; - int16 expDiff; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - bSig = extractFloat64Frac( b ); - bExp = extractFloat64Exp( b ); - expDiff = aExp - bExp; - aSig <<= 9; - bSig <<= 9; - if ( 0 < expDiff ) { - if ( aExp == 0x7FF ) { - if ( aSig ) return propagateFloat64NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig |= LIT64( 0x2000000000000000 ); - } - shift64RightJamming( bSig, expDiff, &bSig ); - zExp = aExp; - } - else if ( expDiff < 0 ) { - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - return packFloat64( zSign, 0x7FF, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig |= LIT64( 0x2000000000000000 ); - } - shift64RightJamming( aSig, - expDiff, &aSig ); - zExp = bExp; - } - else { - if ( aExp == 0x7FF ) { - if ( aSig | bSig ) return propagateFloat64NaN( a, b ); - return a; - } - if ( aExp == 0 ) return packFloat64( zSign, 0, ( aSig + bSig )>>9 ); - zSig = LIT64( 0x4000000000000000 ) + aSig + bSig; - zExp = aExp; - goto roundAndPack; - } - aSig |= LIT64( 0x2000000000000000 ); - zSig = ( aSig + bSig )<<1; - --zExp; - if ( (sbits64) zSig < 0 ) { - zSig = aSig + bSig; - ++zExp; - } - roundAndPack: - return roundAndPackFloat64( zSign, zExp, zSig ); + int16 aExp, bExp, zExp; + bits64 aSig, bSig, zSig; + int16 expDiff; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + bSig = extractFloat64Frac( b ); + bExp = extractFloat64Exp( b ); + expDiff = aExp - bExp; + aSig <<= 9; + bSig <<= 9; + if ( 0 < expDiff ) { + if ( aExp == 0x7FF ) { + if ( aSig ) return propagateFloat64NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig |= LIT64( 0x2000000000000000 ); + } + shift64RightJamming( bSig, expDiff, &bSig ); + zExp = aExp; + } + else if ( expDiff < 0 ) { + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + return packFloat64( zSign, 0x7FF, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig |= LIT64( 0x2000000000000000 ); + } + shift64RightJamming( aSig, - expDiff, &aSig ); + zExp = bExp; + } + else { + if ( aExp == 0x7FF ) { + if ( aSig | bSig ) return propagateFloat64NaN( a, b ); + return a; + } + if ( aExp == 0 ) return packFloat64( zSign, 0, ( aSig + bSig )>>9 ); + zSig = LIT64( 0x4000000000000000 ) + aSig + bSig; + zExp = aExp; + goto roundAndPack; + } + aSig |= LIT64( 0x2000000000000000 ); + zSig = ( aSig + bSig )<<1; + --zExp; + if ( (sbits64) zSig < 0 ) { + zSig = aSig + bSig; + ++zExp; + } + roundAndPack: + return roundAndPackFloat64( zSign, zExp, zSig ); } @@ -2334,68 +2334,68 @@ static float64 addFloat64Sigs( float64 a, float64 b, flag zSign ) static float64 subFloat64Sigs( float64 a, float64 b, flag zSign ) { - int16 aExp, bExp, zExp; - bits64 aSig, bSig, zSig; - int16 expDiff; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - bSig = extractFloat64Frac( b ); - bExp = extractFloat64Exp( b ); - expDiff = aExp - bExp; - aSig <<= 10; - bSig <<= 10; - if ( 0 < expDiff ) goto aExpBigger; - if ( expDiff < 0 ) goto bExpBigger; - if ( aExp == 0x7FF ) { - if ( aSig | bSig ) return propagateFloat64NaN( a, b ); - float_raise( float_flag_invalid ); - return float64_default_nan; - } - if ( aExp == 0 ) { - aExp = 1; - bExp = 1; - } - if ( bSig < aSig ) goto aBigger; - if ( aSig < bSig ) goto bBigger; - return packFloat64( float_rounding_mode == float_round_down, 0, 0 ); - bExpBigger: - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - return packFloat64( zSign ^ 1, 0x7FF, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig |= LIT64( 0x4000000000000000 ); - } - shift64RightJamming( aSig, - expDiff, &aSig ); - bSig |= LIT64( 0x4000000000000000 ); - bBigger: - zSig = bSig - aSig; - zExp = bExp; - zSign ^= 1; - goto normalizeRoundAndPack; - aExpBigger: - if ( aExp == 0x7FF ) { - if ( aSig ) return propagateFloat64NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig |= LIT64( 0x4000000000000000 ); - } - shift64RightJamming( bSig, expDiff, &bSig ); - aSig |= LIT64( 0x4000000000000000 ); - aBigger: - zSig = aSig - bSig; - zExp = aExp; - normalizeRoundAndPack: - --zExp; - return normalizeRoundAndPackFloat64( zSign, zExp, zSig ); + int16 aExp, bExp, zExp; + bits64 aSig, bSig, zSig; + int16 expDiff; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + bSig = extractFloat64Frac( b ); + bExp = extractFloat64Exp( b ); + expDiff = aExp - bExp; + aSig <<= 10; + bSig <<= 10; + if ( 0 < expDiff ) goto aExpBigger; + if ( expDiff < 0 ) goto bExpBigger; + if ( aExp == 0x7FF ) { + if ( aSig | bSig ) return propagateFloat64NaN( a, b ); + float_raise( float_flag_invalid ); + return float64_default_nan; + } + if ( aExp == 0 ) { + aExp = 1; + bExp = 1; + } + if ( bSig < aSig ) goto aBigger; + if ( aSig < bSig ) goto bBigger; + return packFloat64( float_rounding_mode == float_round_down, 0, 0 ); + bExpBigger: + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + return packFloat64( zSign ^ 1, 0x7FF, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig |= LIT64( 0x4000000000000000 ); + } + shift64RightJamming( aSig, - expDiff, &aSig ); + bSig |= LIT64( 0x4000000000000000 ); + bBigger: + zSig = bSig - aSig; + zExp = bExp; + zSign ^= 1; + goto normalizeRoundAndPack; + aExpBigger: + if ( aExp == 0x7FF ) { + if ( aSig ) return propagateFloat64NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig |= LIT64( 0x4000000000000000 ); + } + shift64RightJamming( bSig, expDiff, &bSig ); + aSig |= LIT64( 0x4000000000000000 ); + aBigger: + zSig = aSig - bSig; + zExp = aExp; + normalizeRoundAndPack: + --zExp; + return normalizeRoundAndPackFloat64( zSign, zExp, zSig ); } @@ -2407,16 +2407,16 @@ static float64 subFloat64Sigs( float64 a, float64 b, flag zSign ) float64 float64_add( float64 a, float64 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign == bSign ) { - return addFloat64Sigs( a, b, aSign ); - } - else { - return subFloat64Sigs( a, b, aSign ); - } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign == bSign ) { + return addFloat64Sigs( a, b, aSign ); + } + else { + return subFloat64Sigs( a, b, aSign ); + } } @@ -2428,16 +2428,16 @@ float64 float64_add( float64 a, float64 b ) float64 float64_sub( float64 a, float64 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign == bSign ) { - return subFloat64Sigs( a, b, aSign ); - } - else { - return addFloat64Sigs( a, b, aSign ); - } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign == bSign ) { + return subFloat64Sigs( a, b, aSign ); + } + else { + return addFloat64Sigs( a, b, aSign ); + } } @@ -2449,53 +2449,53 @@ float64 float64_sub( float64 a, float64 b ) float64 float64_mul( float64 a, float64 b ) { - flag aSign, bSign, zSign; - int16 aExp, bExp, zExp; - bits64 aSig, bSig, zSig0, zSig1; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - bSig = extractFloat64Frac( b ); - bExp = extractFloat64Exp( b ); - bSign = extractFloat64Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FF ) { - if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { - return propagateFloat64NaN( a, b ); - } - if ( ( bExp | bSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float64_default_nan; - } - return packFloat64( zSign, 0x7FF, 0 ); - } - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - if ( ( aExp | aSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float64_default_nan; - } - return packFloat64( zSign, 0x7FF, 0 ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) return packFloat64( zSign, 0, 0 ); - normalizeFloat64Subnormal( bSig, &bExp, &bSig ); - } - zExp = aExp + bExp - 0x3FF; - aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; - bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; - mul64To128( aSig, bSig, &zSig0, &zSig1 ); - zSig0 |= ( zSig1 != 0 ); - if ( 0 <= (sbits64) ( zSig0<<1 ) ) { - zSig0 <<= 1; - --zExp; - } - return roundAndPackFloat64( zSign, zExp, zSig0 ); + flag aSign, bSign, zSign; + int16 aExp, bExp, zExp; + bits64 aSig, bSig, zSig0, zSig1; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + bSig = extractFloat64Frac( b ); + bExp = extractFloat64Exp( b ); + bSign = extractFloat64Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FF ) { + if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { + return propagateFloat64NaN( a, b ); + } + if ( ( bExp | bSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float64_default_nan; + } + return packFloat64( zSign, 0x7FF, 0 ); + } + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + if ( ( aExp | aSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float64_default_nan; + } + return packFloat64( zSign, 0x7FF, 0 ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) return packFloat64( zSign, 0, 0 ); + normalizeFloat64Subnormal( bSig, &bExp, &bSig ); + } + zExp = aExp + bExp - 0x3FF; + aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; + bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; + mul64To128( aSig, bSig, &zSig0, &zSig1 ); + zSig0 |= ( zSig1 != 0 ); + if ( 0 <= (sbits64) ( zSig0<<1 ) ) { + zSig0 <<= 1; + --zExp; + } + return roundAndPackFloat64( zSign, zExp, zSig0 ); } @@ -2507,65 +2507,65 @@ float64 float64_mul( float64 a, float64 b ) float64 float64_div( float64 a, float64 b ) { - flag aSign, bSign, zSign; - int16 aExp, bExp, zExp; - bits64 aSig, bSig, zSig; - bits64 rem0, rem1; - bits64 term0, term1; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - bSig = extractFloat64Frac( b ); - bExp = extractFloat64Exp( b ); - bSign = extractFloat64Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FF ) { - if ( aSig ) return propagateFloat64NaN( a, b ); - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - float_raise( float_flag_invalid ); - return float64_default_nan; - } - return packFloat64( zSign, 0x7FF, 0 ); - } - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - return packFloat64( zSign, 0, 0 ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - if ( ( aExp | aSig ) == 0 ) { - float_raise( float_flag_invalid ); - return float64_default_nan; - } - float_raise( float_flag_divbyzero ); - return packFloat64( zSign, 0x7FF, 0 ); - } - normalizeFloat64Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - } - zExp = aExp - bExp + 0x3FD; - aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; - bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; - if ( bSig <= ( aSig + aSig ) ) { - aSig >>= 1; - ++zExp; - } - zSig = estimateDiv128To64( aSig, 0, bSig ); - if ( ( zSig & 0x1FF ) <= 2 ) { - mul64To128( bSig, zSig, &term0, &term1 ); - sub128( aSig, 0, term0, term1, &rem0, &rem1 ); - while ( (sbits64) rem0 < 0 ) { - --zSig; - add128( rem0, rem1, 0, bSig, &rem0, &rem1 ); - } - zSig |= ( rem1 != 0 ); - } - return roundAndPackFloat64( zSign, zExp, zSig ); + flag aSign, bSign, zSign; + int16 aExp, bExp, zExp; + bits64 aSig, bSig, zSig; + bits64 rem0, rem1; + bits64 term0, term1; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + bSig = extractFloat64Frac( b ); + bExp = extractFloat64Exp( b ); + bSign = extractFloat64Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FF ) { + if ( aSig ) return propagateFloat64NaN( a, b ); + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + float_raise( float_flag_invalid ); + return float64_default_nan; + } + return packFloat64( zSign, 0x7FF, 0 ); + } + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + return packFloat64( zSign, 0, 0 ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + if ( ( aExp | aSig ) == 0 ) { + float_raise( float_flag_invalid ); + return float64_default_nan; + } + float_raise( float_flag_divbyzero ); + return packFloat64( zSign, 0x7FF, 0 ); + } + normalizeFloat64Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloat64( zSign, 0, 0 ); + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + } + zExp = aExp - bExp + 0x3FD; + aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<10; + bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; + if ( bSig <= ( aSig + aSig ) ) { + aSig >>= 1; + ++zExp; + } + zSig = estimateDiv128To64( aSig, 0, bSig ); + if ( ( zSig & 0x1FF ) <= 2 ) { + mul64To128( bSig, zSig, &term0, &term1 ); + sub128( aSig, 0, term0, term1, &rem0, &rem1 ); + while ( (sbits64) rem0 < 0 ) { + --zSig; + add128( rem0, rem1, 0, bSig, &rem0, &rem1 ); + } + zSig |= ( rem1 != 0 ); + } + return roundAndPackFloat64( zSign, zExp, zSig ); } @@ -2577,80 +2577,80 @@ float64 float64_div( float64 a, float64 b ) float64 float64_rem( float64 a, float64 b ) { - flag aSign, zSign; - int16 aExp, bExp, expDiff; - bits64 aSig, bSig; - bits64 q, alternateASig; - sbits64 sigMean; - - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - bSig = extractFloat64Frac( b ); - bExp = extractFloat64Exp( b ); + flag aSign, zSign; + int16 aExp, bExp, expDiff; + bits64 aSig, bSig; + bits64 q, alternateASig; + sbits64 sigMean; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + bSig = extractFloat64Frac( b ); + bExp = extractFloat64Exp( b ); // bSign = extractFloat64Sign( b ); - if ( aExp == 0x7FF ) { - if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { - return propagateFloat64NaN( a, b ); - } - float_raise( float_flag_invalid ); - return float64_default_nan; - } - if ( bExp == 0x7FF ) { - if ( bSig ) return propagateFloat64NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - float_raise( float_flag_invalid ); - return float64_default_nan; - } - normalizeFloat64Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return a; - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - } - expDiff = aExp - bExp; - aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<11; - bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; - if ( expDiff < 0 ) { - if ( expDiff < -1 ) return a; - aSig >>= 1; - } - q = ( bSig <= aSig ); - if ( q ) aSig -= bSig; - expDiff -= 64; - while ( 0 < expDiff ) { - q = estimateDiv128To64( aSig, 0, bSig ); - q = ( 2 < q ) ? q - 2 : 0; - aSig = - ( ( bSig>>2 ) * q ); - expDiff -= 62; - } - expDiff += 64; - if ( 0 < expDiff ) { - q = estimateDiv128To64( aSig, 0, bSig ); - q = ( 2 < q ) ? q - 2 : 0; - q >>= 64 - expDiff; - bSig >>= 2; - aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; - } - else { - aSig >>= 2; - bSig >>= 2; - } - do { - alternateASig = aSig; - ++q; - aSig -= bSig; - } while ( 0 <= (sbits64) aSig ); - sigMean = aSig + alternateASig; - if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { - aSig = alternateASig; - } - zSign = ( (sbits64) aSig < 0 ); - if ( zSign ) aSig = - aSig; - return normalizeRoundAndPackFloat64( aSign ^ zSign, bExp, aSig ); + if ( aExp == 0x7FF ) { + if ( aSig || ( ( bExp == 0x7FF ) && bSig ) ) { + return propagateFloat64NaN( a, b ); + } + float_raise( float_flag_invalid ); + return float64_default_nan; + } + if ( bExp == 0x7FF ) { + if ( bSig ) return propagateFloat64NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + float_raise( float_flag_invalid ); + return float64_default_nan; + } + normalizeFloat64Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return a; + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + } + expDiff = aExp - bExp; + aSig = ( aSig | LIT64( 0x0010000000000000 ) )<<11; + bSig = ( bSig | LIT64( 0x0010000000000000 ) )<<11; + if ( expDiff < 0 ) { + if ( expDiff < -1 ) return a; + aSig >>= 1; + } + q = ( bSig <= aSig ); + if ( q ) aSig -= bSig; + expDiff -= 64; + while ( 0 < expDiff ) { + q = estimateDiv128To64( aSig, 0, bSig ); + q = ( 2 < q ) ? q - 2 : 0; + aSig = - ( ( bSig>>2 ) * q ); + expDiff -= 62; + } + expDiff += 64; + if ( 0 < expDiff ) { + q = estimateDiv128To64( aSig, 0, bSig ); + q = ( 2 < q ) ? q - 2 : 0; + q >>= 64 - expDiff; + bSig >>= 2; + aSig = ( ( aSig>>1 )<<( expDiff - 1 ) ) - bSig * q; + } + else { + aSig >>= 2; + bSig >>= 2; + } + do { + alternateASig = aSig; + ++q; + aSig -= bSig; + } while ( 0 <= (sbits64) aSig ); + sigMean = aSig + alternateASig; + if ( ( sigMean < 0 ) || ( ( sigMean == 0 ) && ( q & 1 ) ) ) { + aSig = alternateASig; + } + zSign = ( (sbits64) aSig < 0 ); + if ( zSign ) aSig = - aSig; + return normalizeRoundAndPackFloat64( aSign ^ zSign, bExp, aSig ); } @@ -2662,47 +2662,47 @@ float64 float64_rem( float64 a, float64 b ) float64 float64_sqrt( float64 a ) { - flag aSign; - int16 aExp, zExp; - bits64 aSig, zSig, doubleZSig; - bits64 rem0, rem1, term0, term1; + flag aSign; + int16 aExp, zExp; + bits64 aSig, zSig, doubleZSig; + bits64 rem0, rem1, term0, term1; // float64 z; - aSig = extractFloat64Frac( a ); - aExp = extractFloat64Exp( a ); - aSign = extractFloat64Sign( a ); - if ( aExp == 0x7FF ) { - if ( aSig ) return propagateFloat64NaN( a, a ); - if ( ! aSign ) return a; - float_raise( float_flag_invalid ); - return float64_default_nan; - } - if ( aSign ) { - if ( ( aExp | aSig ) == 0 ) return a; - float_raise( float_flag_invalid ); - return float64_default_nan; - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return 0; - normalizeFloat64Subnormal( aSig, &aExp, &aSig ); - } - zExp = ( ( aExp - 0x3FF )>>1 ) + 0x3FE; - aSig |= LIT64( 0x0010000000000000 ); - zSig = estimateSqrt32( aExp, aSig>>21 ); - aSig <<= 9 - ( aExp & 1 ); - zSig = estimateDiv128To64( aSig, 0, zSig<<32 ) + ( zSig<<30 ); - if ( ( zSig & 0x1FF ) <= 5 ) { - doubleZSig = zSig<<1; - mul64To128( zSig, zSig, &term0, &term1 ); - sub128( aSig, 0, term0, term1, &rem0, &rem1 ); - while ( (sbits64) rem0 < 0 ) { - --zSig; - doubleZSig -= 2; - add128( rem0, rem1, zSig>>63, doubleZSig | 1, &rem0, &rem1 ); - } - zSig |= ( ( rem0 | rem1 ) != 0 ); - } - return roundAndPackFloat64( 0, zExp, zSig ); + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + if ( aExp == 0x7FF ) { + if ( aSig ) return propagateFloat64NaN( a, a ); + if ( ! aSign ) return a; + float_raise( float_flag_invalid ); + return float64_default_nan; + } + if ( aSign ) { + if ( ( aExp | aSig ) == 0 ) return a; + float_raise( float_flag_invalid ); + return float64_default_nan; + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return 0; + normalizeFloat64Subnormal( aSig, &aExp, &aSig ); + } + zExp = ( ( aExp - 0x3FF )>>1 ) + 0x3FE; + aSig |= LIT64( 0x0010000000000000 ); + zSig = estimateSqrt32( aExp, aSig>>21 ); + aSig <<= 9 - ( aExp & 1 ); + zSig = estimateDiv128To64( aSig, 0, zSig<<32 ) + ( zSig<<30 ); + if ( ( zSig & 0x1FF ) <= 5 ) { + doubleZSig = zSig<<1; + mul64To128( zSig, zSig, &term0, &term1 ); + sub128( aSig, 0, term0, term1, &rem0, &rem1 ); + while ( (sbits64) rem0 < 0 ) { + --zSig; + doubleZSig -= 2; + add128( rem0, rem1, zSig>>63, doubleZSig | 1, &rem0, &rem1 ); + } + zSig |= ( ( rem0 | rem1 ) != 0 ); + } + return roundAndPackFloat64( 0, zExp, zSig ); } @@ -2715,15 +2715,15 @@ float64 float64_sqrt( float64 a ) flag float64_eq( float64 a, float64 b ) { - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - return ( a == b ) || ( (bits64) ( ( a | b )<<1 ) == 0 ); + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + return ( a == b ) || ( (bits64) ( ( a | b )<<1 ) == 0 ); } @@ -2736,18 +2736,18 @@ flag float64_eq( float64 a, float64 b ) flag float64_le( float64 a, float64 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign != bSign ) return aSign || ( (bits64) ( ( a | b )<<1 ) == 0 ); - return ( a == b ) || ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign != bSign ) return aSign || ( (bits64) ( ( a | b )<<1 ) == 0 ); + return ( a == b ) || ( aSign ^ ( a < b ) ); } @@ -2759,18 +2759,18 @@ flag float64_le( float64 a, float64 b ) flag float64_lt( float64 a, float64 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign != bSign ) return aSign && ( (bits64) ( ( a | b )<<1 ) != 0 ); - return ( a != b ) && ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign != bSign ) return aSign && ( (bits64) ( ( a | b )<<1 ) != 0 ); + return ( a != b ) && ( aSign ^ ( a < b ) ); } @@ -2784,13 +2784,13 @@ flag float64_lt( float64 a, float64 b ) flag float64_eq_signaling( float64 a, float64 b ) { - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - return ( a == b ) || ( (bits64) ( ( a | b )<<1 ) == 0 ); + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + return ( a == b ) || ( (bits64) ( ( a | b )<<1 ) == 0 ); } @@ -2803,21 +2803,21 @@ flag float64_eq_signaling( float64 a, float64 b ) flag float64_le_quiet( float64 a, float64 b ) { - flag aSign, bSign; + flag aSign, bSign; // int16 aExp, bExp; - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign != bSign ) return aSign || ( (bits64) ( ( a | b )<<1 ) == 0 ); - return ( a == b ) || ( aSign ^ ( a < b ) ); + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign != bSign ) return aSign || ( (bits64) ( ( a | b )<<1 ) == 0 ); + return ( a == b ) || ( aSign ^ ( a < b ) ); } @@ -2830,20 +2830,20 @@ flag float64_le_quiet( float64 a, float64 b ) flag float64_lt_quiet( float64 a, float64 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) - || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) - ) { - if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat64Sign( a ); - bSign = extractFloat64Sign( b ); - if ( aSign != bSign ) return aSign && ( (bits64) ( ( a | b )<<1 ) != 0 ); - return ( a != b ) && ( aSign ^ ( a < b ) ); + flag aSign, bSign; + + if ( ( ( extractFloat64Exp( a ) == 0x7FF ) && extractFloat64Frac( a ) ) + || ( ( extractFloat64Exp( b ) == 0x7FF ) && extractFloat64Frac( b ) ) + ) { + if ( float64_is_signaling_nan( a ) || float64_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat64Sign( a ); + bSign = extractFloat64Sign( b ); + if ( aSign != bSign ) return aSign && ( (bits64) ( ( a | b )<<1 ) != 0 ); + return ( a != b ) && ( aSign ^ ( a < b ) ); } @@ -2861,18 +2861,18 @@ flag float64_lt_quiet( float64 a, float64 b ) int32 floatx80_to_int32( floatx80 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; - shiftCount = 0x4037 - aExp; - if ( shiftCount <= 0 ) shiftCount = 1; - shift64RightJamming( aSig, shiftCount, &aSig ); - return roundAndPackInt32( aSign, aSig ); + flag aSign; + int32 aExp, shiftCount; + bits64 aSig; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; + shiftCount = 0x4037 - aExp; + if ( shiftCount <= 0 ) shiftCount = 1; + shift64RightJamming( aSig, shiftCount, &aSig ); + return roundAndPackInt32( aSign, aSig ); } @@ -2888,36 +2888,36 @@ int32 floatx80_to_int32( floatx80 a ) int32 floatx80_to_int32_round_to_zero( floatx80 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig, savedASig; - int32 z; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( 0x401E < aExp ) { - if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; - goto invalid; - } - else if ( aExp < 0x3FFF ) { - if ( aExp || aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - shiftCount = 0x403E - aExp; - savedASig = aSig; - aSig >>= shiftCount; - z = aSig; - if ( aSign ) z = - z; - if ( ( z < 0 ) ^ aSign ) { - invalid: - float_raise( float_flag_invalid ); - return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; - } - if ( ( aSig<<shiftCount ) != savedASig ) { - float_exception_flags |= float_flag_inexact; - } - return z; + flag aSign; + int32 aExp, shiftCount; + bits64 aSig, savedASig; + int32 z; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( 0x401E < aExp ) { + if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) aSign = 0; + goto invalid; + } + else if ( aExp < 0x3FFF ) { + if ( aExp || aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + shiftCount = 0x403E - aExp; + savedASig = aSig; + aSig >>= shiftCount; + z = aSig; + if ( aSign ) z = - z; + if ( ( z < 0 ) ^ aSign ) { + invalid: + float_raise( float_flag_invalid ); + return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; + } + if ( ( aSig<<shiftCount ) != savedASig ) { + float_exception_flags |= float_flag_inexact; + } + return z; } @@ -2933,31 +2933,31 @@ int32 floatx80_to_int32_round_to_zero( floatx80 a ) int64 floatx80_to_int64( floatx80 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig, aSigExtra; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - shiftCount = 0x403E - aExp; - if ( shiftCount <= 0 ) { - if ( shiftCount ) { - float_raise( float_flag_invalid ); - if ( ! aSign - || ( ( aExp == 0x7FFF ) - && ( aSig != LIT64( 0x8000000000000000 ) ) ) - ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - aSigExtra = 0; - } - else { - shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra ); - } - return roundAndPackInt64( aSign, aSig, aSigExtra ); + flag aSign; + int32 aExp, shiftCount; + bits64 aSig, aSigExtra; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + shiftCount = 0x403E - aExp; + if ( shiftCount <= 0 ) { + if ( shiftCount ) { + float_raise( float_flag_invalid ); + if ( ! aSign + || ( ( aExp == 0x7FFF ) + && ( aSig != LIT64( 0x8000000000000000 ) ) ) + ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + aSigExtra = 0; + } + else { + shift64ExtraRightJamming( aSig, 0, shiftCount, &aSig, &aSigExtra ); + } + return roundAndPackInt64( aSign, aSig, aSigExtra ); } @@ -2973,35 +2973,35 @@ int64 floatx80_to_int64( floatx80 a ) int64 floatx80_to_int64_round_to_zero( floatx80 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig; - int64 z; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - shiftCount = aExp - 0x403E; - if ( 0 <= shiftCount ) { - aSig &= LIT64( 0x7FFFFFFFFFFFFFFF ); - if ( ( a.high != 0xC03E ) || aSig ) { - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0x7FFF ) && aSig ) ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - else if ( aExp < 0x3FFF ) { - if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; - return 0; - } - z = aSig>>( - shiftCount ); - if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { - float_exception_flags |= float_flag_inexact; - } - if ( aSign ) z = - z; - return z; + flag aSign; + int32 aExp, shiftCount; + bits64 aSig; + int64 z; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + shiftCount = aExp - 0x403E; + if ( 0 <= shiftCount ) { + aSig &= LIT64( 0x7FFFFFFFFFFFFFFF ); + if ( ( a.high != 0xC03E ) || aSig ) { + float_raise( float_flag_invalid ); + if ( ! aSign || ( ( aExp == 0x7FFF ) && aSig ) ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + else if ( aExp < 0x3FFF ) { + if ( aExp | aSig ) float_exception_flags |= float_flag_inexact; + return 0; + } + z = aSig>>( - shiftCount ); + if ( (bits64) ( aSig<<( shiftCount & 63 ) ) ) { + float_exception_flags |= float_flag_inexact; + } + if ( aSign ) z = - z; + return z; } @@ -3014,22 +3014,22 @@ int64 floatx80_to_int64_round_to_zero( floatx80 a ) float32 floatx80_to_float32( floatx80 a ) { - flag aSign; - int32 aExp; - bits64 aSig; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) ) { - return commonNaNToFloat32( floatx80ToCommonNaN( a ) ); - } - return packFloat32( aSign, 0xFF, 0 ); - } - shift64RightJamming( aSig, 33, &aSig ); - if ( aExp || aSig ) aExp -= 0x3F81; - return roundAndPackFloat32( aSign, aExp, aSig ); + flag aSign; + int32 aExp; + bits64 aSig; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) ) { + return commonNaNToFloat32( floatx80ToCommonNaN( a ) ); + } + return packFloat32( aSign, 0xFF, 0 ); + } + shift64RightJamming( aSig, 33, &aSig ); + if ( aExp || aSig ) aExp -= 0x3F81; + return roundAndPackFloat32( aSign, aExp, aSig ); } @@ -3042,22 +3042,22 @@ float32 floatx80_to_float32( floatx80 a ) float64 floatx80_to_float64( floatx80 a ) { - flag aSign; - int32 aExp; - bits64 aSig, zSig; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) ) { - return commonNaNToFloat64( floatx80ToCommonNaN( a ) ); - } - return packFloat64( aSign, 0x7FF, 0 ); - } - shift64RightJamming( aSig, 1, &zSig ); - if ( aExp || aSig ) aExp -= 0x3C01; - return roundAndPackFloat64( aSign, aExp, zSig ); + flag aSign; + int32 aExp; + bits64 aSig, zSig; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) ) { + return commonNaNToFloat64( floatx80ToCommonNaN( a ) ); + } + return packFloat64( aSign, 0x7FF, 0 ); + } + shift64RightJamming( aSig, 1, &zSig ); + if ( aExp || aSig ) aExp -= 0x3C01; + return roundAndPackFloat64( aSign, aExp, zSig ); } @@ -3072,18 +3072,18 @@ float64 floatx80_to_float64( floatx80 a ) float128 floatx80_to_float128( floatx80 a ) { - flag aSign; - int16 aExp; - bits64 aSig, zSig0, zSig1; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) { - return commonNaNToFloat128( floatx80ToCommonNaN( a ) ); - } - shift128Right( aSig<<1, 0, 16, &zSig0, &zSig1 ); - return packFloat128( aSign, aExp, zSig0, zSig1 ); + flag aSign; + int16 aExp; + bits64 aSig, zSig0, zSig1; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( ( aExp == 0x7FFF ) && (bits64) ( aSig<<1 ) ) { + return commonNaNToFloat128( floatx80ToCommonNaN( a ) ); + } + shift128Right( aSig<<1, 0, 16, &zSig0, &zSig1 ); + return packFloat128( aSign, aExp, zSig0, zSig1 ); } @@ -3098,67 +3098,67 @@ float128 floatx80_to_float128( floatx80 a ) floatx80 floatx80_round_to_int( floatx80 a ) { - flag aSign; - int32 aExp; - bits64 lastBitMask, roundBitsMask; - int8 roundingMode; - floatx80 z; - - aExp = extractFloatx80Exp( a ); - if ( 0x403E <= aExp ) { - if ( ( aExp == 0x7FFF ) && (bits64) ( extractFloatx80Frac( a )<<1 ) ) { - return propagateFloatx80NaN( a, a ); - } - return a; - } - if ( aExp < 0x3FFF ) { - if ( ( aExp == 0 ) - && ( (bits64) ( extractFloatx80Frac( a )<<1 ) == 0 ) ) { - return a; - } - float_exception_flags |= float_flag_inexact; - aSign = extractFloatx80Sign( a ); - switch ( float_rounding_mode ) { - case float_round_nearest_even: - if ( ( aExp == 0x3FFE ) && (bits64) ( extractFloatx80Frac( a )<<1 ) - ) { - return - packFloatx80( aSign, 0x3FFF, LIT64( 0x8000000000000000 ) ); - } - break; - case float_round_down: - return - aSign ? - packFloatx80( 1, 0x3FFF, LIT64( 0x8000000000000000 ) ) - : packFloatx80( 0, 0, 0 ); - case float_round_up: - return - aSign ? packFloatx80( 1, 0, 0 ) - : packFloatx80( 0, 0x3FFF, LIT64( 0x8000000000000000 ) ); - } - return packFloatx80( aSign, 0, 0 ); - } - lastBitMask = 1; - lastBitMask <<= 0x403E - aExp; - roundBitsMask = lastBitMask - 1; - z = a; - roundingMode = float_rounding_mode; - if ( roundingMode == float_round_nearest_even ) { - z.low += lastBitMask>>1; - if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; - } - else if ( roundingMode != float_round_to_zero ) { - if ( extractFloatx80Sign( z ) ^ ( roundingMode == float_round_up ) ) { - z.low += roundBitsMask; - } - } - z.low &= ~ roundBitsMask; - if ( z.low == 0 ) { - ++z.high; - z.low = LIT64( 0x8000000000000000 ); - } - if ( z.low != a.low ) float_exception_flags |= float_flag_inexact; - return z; + flag aSign; + int32 aExp; + bits64 lastBitMask, roundBitsMask; + int8 roundingMode; + floatx80 z; + + aExp = extractFloatx80Exp( a ); + if ( 0x403E <= aExp ) { + if ( ( aExp == 0x7FFF ) && (bits64) ( extractFloatx80Frac( a )<<1 ) ) { + return propagateFloatx80NaN( a, a ); + } + return a; + } + if ( aExp < 0x3FFF ) { + if ( ( aExp == 0 ) + && ( (bits64) ( extractFloatx80Frac( a )<<1 ) == 0 ) ) { + return a; + } + float_exception_flags |= float_flag_inexact; + aSign = extractFloatx80Sign( a ); + switch ( float_rounding_mode ) { + case float_round_nearest_even: + if ( ( aExp == 0x3FFE ) && (bits64) ( extractFloatx80Frac( a )<<1 ) + ) { + return + packFloatx80( aSign, 0x3FFF, LIT64( 0x8000000000000000 ) ); + } + break; + case float_round_down: + return + aSign ? + packFloatx80( 1, 0x3FFF, LIT64( 0x8000000000000000 ) ) + : packFloatx80( 0, 0, 0 ); + case float_round_up: + return + aSign ? packFloatx80( 1, 0, 0 ) + : packFloatx80( 0, 0x3FFF, LIT64( 0x8000000000000000 ) ); + } + return packFloatx80( aSign, 0, 0 ); + } + lastBitMask = 1; + lastBitMask <<= 0x403E - aExp; + roundBitsMask = lastBitMask - 1; + z = a; + roundingMode = float_rounding_mode; + if ( roundingMode == float_round_nearest_even ) { + z.low += lastBitMask>>1; + if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; + } + else if ( roundingMode != float_round_to_zero ) { + if ( extractFloatx80Sign( z ) ^ ( roundingMode == float_round_up ) ) { + z.low += roundBitsMask; + } + } + z.low &= ~ roundBitsMask; + if ( z.low == 0 ) { + ++z.high; + z.low = LIT64( 0x8000000000000000 ); + } + if ( z.low != a.low ) float_exception_flags |= float_flag_inexact; + return z; } @@ -3172,59 +3172,59 @@ floatx80 floatx80_round_to_int( floatx80 a ) static floatx80 addFloatx80Sigs( floatx80 a, floatx80 b, flag zSign ) { - int32 aExp, bExp, zExp; - bits64 aSig, bSig, zSig0, zSig1; - int32 expDiff; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - bSig = extractFloatx80Frac( b ); - bExp = extractFloatx80Exp( b ); - expDiff = aExp - bExp; - if ( 0 < expDiff ) { - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return a; - } - if ( bExp == 0 ) --expDiff; - shift64ExtraRightJamming( bSig, 0, expDiff, &bSig, &zSig1 ); - zExp = aExp; - } - else if ( expDiff < 0 ) { - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) ++expDiff; - shift64ExtraRightJamming( aSig, 0, - expDiff, &aSig, &zSig1 ); - zExp = bExp; - } - else { - if ( aExp == 0x7FFF ) { - if ( (bits64) ( ( aSig | bSig )<<1 ) ) { - return propagateFloatx80NaN( a, b ); - } - return a; - } - zSig1 = 0; - zSig0 = aSig + bSig; - if ( aExp == 0 ) { - normalizeFloatx80Subnormal( zSig0, &zExp, &zSig0 ); - goto roundAndPack; - } - zExp = aExp; - goto shiftRight1; - } - zSig0 = aSig + bSig; - if ( (sbits64) zSig0 < 0 ) goto roundAndPack; - shiftRight1: - shift64ExtraRightJamming( zSig0, zSig1, 1, &zSig0, &zSig1 ); - zSig0 |= LIT64( 0x8000000000000000 ); - ++zExp; - roundAndPack: - return - roundAndPackFloatx80( - floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); + int32 aExp, bExp, zExp; + bits64 aSig, bSig, zSig0, zSig1; + int32 expDiff; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + bSig = extractFloatx80Frac( b ); + bExp = extractFloatx80Exp( b ); + expDiff = aExp - bExp; + if ( 0 < expDiff ) { + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return a; + } + if ( bExp == 0 ) --expDiff; + shift64ExtraRightJamming( bSig, 0, expDiff, &bSig, &zSig1 ); + zExp = aExp; + } + else if ( expDiff < 0 ) { + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) ++expDiff; + shift64ExtraRightJamming( aSig, 0, - expDiff, &aSig, &zSig1 ); + zExp = bExp; + } + else { + if ( aExp == 0x7FFF ) { + if ( (bits64) ( ( aSig | bSig )<<1 ) ) { + return propagateFloatx80NaN( a, b ); + } + return a; + } + zSig1 = 0; + zSig0 = aSig + bSig; + if ( aExp == 0 ) { + normalizeFloatx80Subnormal( zSig0, &zExp, &zSig0 ); + goto roundAndPack; + } + zExp = aExp; + goto shiftRight1; + } + zSig0 = aSig + bSig; + if ( (sbits64) zSig0 < 0 ) goto roundAndPack; + shiftRight1: + shift64ExtraRightJamming( zSig0, zSig1, 1, &zSig0, &zSig1 ); + zSig0 |= LIT64( 0x8000000000000000 ); + ++zExp; + roundAndPack: + return + roundAndPackFloatx80( + floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); } @@ -3238,61 +3238,61 @@ static floatx80 addFloatx80Sigs( floatx80 a, floatx80 b, flag zSign ) static floatx80 subFloatx80Sigs( floatx80 a, floatx80 b, flag zSign ) { - int32 aExp, bExp, zExp; - bits64 aSig, bSig, zSig0, zSig1; - int32 expDiff; - floatx80 z; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - bSig = extractFloatx80Frac( b ); - bExp = extractFloatx80Exp( b ); - expDiff = aExp - bExp; - if ( 0 < expDiff ) goto aExpBigger; - if ( expDiff < 0 ) goto bExpBigger; - if ( aExp == 0x7FFF ) { - if ( (bits64) ( ( aSig | bSig )<<1 ) ) { - return propagateFloatx80NaN( a, b ); - } - float_raise( float_flag_invalid ); - z.low = floatx80_default_nan_low; - z.high = floatx80_default_nan_high; - return z; - } - if ( aExp == 0 ) { - aExp = 1; - bExp = 1; - } - zSig1 = 0; - if ( bSig < aSig ) goto aBigger; - if ( aSig < bSig ) goto bBigger; - return packFloatx80( float_rounding_mode == float_round_down, 0, 0 ); - bExpBigger: - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return packFloatx80( zSign ^ 1, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) ++expDiff; - shift128RightJamming( aSig, 0, - expDiff, &aSig, &zSig1 ); - bBigger: - sub128( bSig, 0, aSig, zSig1, &zSig0, &zSig1 ); - zExp = bExp; - zSign ^= 1; - goto normalizeRoundAndPack; - aExpBigger: - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return a; - } - if ( bExp == 0 ) --expDiff; - shift128RightJamming( bSig, 0, expDiff, &bSig, &zSig1 ); - aBigger: - sub128( aSig, 0, bSig, zSig1, &zSig0, &zSig1 ); - zExp = aExp; - normalizeRoundAndPack: - return - normalizeRoundAndPackFloatx80( - floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); + int32 aExp, bExp, zExp; + bits64 aSig, bSig, zSig0, zSig1; + int32 expDiff; + floatx80 z; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + bSig = extractFloatx80Frac( b ); + bExp = extractFloatx80Exp( b ); + expDiff = aExp - bExp; + if ( 0 < expDiff ) goto aExpBigger; + if ( expDiff < 0 ) goto bExpBigger; + if ( aExp == 0x7FFF ) { + if ( (bits64) ( ( aSig | bSig )<<1 ) ) { + return propagateFloatx80NaN( a, b ); + } + float_raise( float_flag_invalid ); + z.low = floatx80_default_nan_low; + z.high = floatx80_default_nan_high; + return z; + } + if ( aExp == 0 ) { + aExp = 1; + bExp = 1; + } + zSig1 = 0; + if ( bSig < aSig ) goto aBigger; + if ( aSig < bSig ) goto bBigger; + return packFloatx80( float_rounding_mode == float_round_down, 0, 0 ); + bExpBigger: + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return packFloatx80( zSign ^ 1, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) ++expDiff; + shift128RightJamming( aSig, 0, - expDiff, &aSig, &zSig1 ); + bBigger: + sub128( bSig, 0, aSig, zSig1, &zSig0, &zSig1 ); + zExp = bExp; + zSign ^= 1; + goto normalizeRoundAndPack; + aExpBigger: + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return a; + } + if ( bExp == 0 ) --expDiff; + shift128RightJamming( bSig, 0, expDiff, &bSig, &zSig1 ); + aBigger: + sub128( aSig, 0, bSig, zSig1, &zSig0, &zSig1 ); + zExp = aExp; + normalizeRoundAndPack: + return + normalizeRoundAndPackFloatx80( + floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); } @@ -3304,16 +3304,16 @@ static floatx80 subFloatx80Sigs( floatx80 a, floatx80 b, flag zSign ) floatx80 floatx80_add( floatx80 a, floatx80 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign == bSign ) { - return addFloatx80Sigs( a, b, aSign ); - } - else { - return subFloatx80Sigs( a, b, aSign ); - } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign == bSign ) { + return addFloatx80Sigs( a, b, aSign ); + } + else { + return subFloatx80Sigs( a, b, aSign ); + } } @@ -3325,16 +3325,16 @@ floatx80 floatx80_add( floatx80 a, floatx80 b ) floatx80 floatx80_sub( floatx80 a, floatx80 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign == bSign ) { - return subFloatx80Sigs( a, b, aSign ); - } - else { - return addFloatx80Sigs( a, b, aSign ); - } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign == bSign ) { + return subFloatx80Sigs( a, b, aSign ); + } + else { + return addFloatx80Sigs( a, b, aSign ); + } } @@ -3346,54 +3346,54 @@ floatx80 floatx80_sub( floatx80 a, floatx80 b ) floatx80 floatx80_mul( floatx80 a, floatx80 b ) { - flag aSign, bSign, zSign; - int32 aExp, bExp, zExp; - bits64 aSig, bSig, zSig0, zSig1; - floatx80 z; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - bSig = extractFloatx80Frac( b ); - bExp = extractFloatx80Exp( b ); - bSign = extractFloatx80Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) - || ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { - return propagateFloatx80NaN( a, b ); - } - if ( ( bExp | bSig ) == 0 ) goto invalid; - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - if ( ( aExp | aSig ) == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = floatx80_default_nan_low; - z.high = floatx80_default_nan_high; - return z; - } - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); - normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) return packFloatx80( zSign, 0, 0 ); - normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); - } - zExp = aExp + bExp - 0x3FFE; - mul64To128( aSig, bSig, &zSig0, &zSig1 ); - if ( 0 < (sbits64) zSig0 ) { - shortShift128Left( zSig0, zSig1, 1, &zSig0, &zSig1 ); - --zExp; - } - return - roundAndPackFloatx80( - floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); + flag aSign, bSign, zSign; + int32 aExp, bExp, zExp; + bits64 aSig, bSig, zSig0, zSig1; + floatx80 z; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + bSig = extractFloatx80Frac( b ); + bExp = extractFloatx80Exp( b ); + bSign = extractFloatx80Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) + || ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { + return propagateFloatx80NaN( a, b ); + } + if ( ( bExp | bSig ) == 0 ) goto invalid; + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + if ( ( aExp | aSig ) == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = floatx80_default_nan_low; + z.high = floatx80_default_nan_high; + return z; + } + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); + normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) return packFloatx80( zSign, 0, 0 ); + normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); + } + zExp = aExp + bExp - 0x3FFE; + mul64To128( aSig, bSig, &zSig0, &zSig1 ); + if ( 0 < (sbits64) zSig0 ) { + shortShift128Left( zSig0, zSig1, 1, &zSig0, &zSig1 ); + --zExp; + } + return + roundAndPackFloatx80( + floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); } @@ -3405,75 +3405,75 @@ floatx80 floatx80_mul( floatx80 a, floatx80 b ) floatx80 floatx80_div( floatx80 a, floatx80 b ) { - flag aSign, bSign, zSign; - int32 aExp, bExp, zExp; - bits64 aSig, bSig, zSig0, zSig1; - bits64 rem0, rem1, rem2, term0, term1, term2; - floatx80 z; - - aSig = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - bSig = extractFloatx80Frac( b ); - bExp = extractFloatx80Exp( b ); - bSign = extractFloatx80Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - goto invalid; - } - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return packFloatx80( zSign, 0, 0 ); - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - if ( ( aExp | aSig ) == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = floatx80_default_nan_low; - z.high = floatx80_default_nan_high; - return z; - } - float_raise( float_flag_divbyzero ); - return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); - normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); - } - zExp = aExp - bExp + 0x3FFE; - rem1 = 0; - if ( bSig <= aSig ) { - shift128Right( aSig, 0, 1, &aSig, &rem1 ); - ++zExp; - } - zSig0 = estimateDiv128To64( aSig, rem1, bSig ); - mul64To128( bSig, zSig0, &term0, &term1 ); - sub128( aSig, rem1, term0, term1, &rem0, &rem1 ); - while ( (sbits64) rem0 < 0 ) { - --zSig0; - add128( rem0, rem1, 0, bSig, &rem0, &rem1 ); - } - zSig1 = estimateDiv128To64( rem1, 0, bSig ); - if ( (bits64) ( zSig1<<1 ) <= 8 ) { - mul64To128( bSig, zSig1, &term1, &term2 ); - sub128( rem1, 0, term1, term2, &rem1, &rem2 ); - while ( (sbits64) rem1 < 0 ) { - --zSig1; - add128( rem1, rem2, 0, bSig, &rem1, &rem2 ); - } - zSig1 |= ( ( rem1 | rem2 ) != 0 ); - } - return - roundAndPackFloatx80( - floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); + flag aSign, bSign, zSign; + int32 aExp, bExp, zExp; + bits64 aSig, bSig, zSig0, zSig1; + bits64 rem0, rem1, rem2, term0, term1, term2; + floatx80 z; + + aSig = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + bSig = extractFloatx80Frac( b ); + bExp = extractFloatx80Exp( b ); + bSign = extractFloatx80Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig<<1 ) ) return propagateFloatx80NaN( a, b ); + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + goto invalid; + } + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return packFloatx80( zSign, 0, 0 ); + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + if ( ( aExp | aSig ) == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = floatx80_default_nan_low; + z.high = floatx80_default_nan_high; + return z; + } + float_raise( float_flag_divbyzero ); + return packFloatx80( zSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( aSig == 0 ) return packFloatx80( zSign, 0, 0 ); + normalizeFloatx80Subnormal( aSig, &aExp, &aSig ); + } + zExp = aExp - bExp + 0x3FFE; + rem1 = 0; + if ( bSig <= aSig ) { + shift128Right( aSig, 0, 1, &aSig, &rem1 ); + ++zExp; + } + zSig0 = estimateDiv128To64( aSig, rem1, bSig ); + mul64To128( bSig, zSig0, &term0, &term1 ); + sub128( aSig, rem1, term0, term1, &rem0, &rem1 ); + while ( (sbits64) rem0 < 0 ) { + --zSig0; + add128( rem0, rem1, 0, bSig, &rem0, &rem1 ); + } + zSig1 = estimateDiv128To64( rem1, 0, bSig ); + if ( (bits64) ( zSig1<<1 ) <= 8 ) { + mul64To128( bSig, zSig1, &term1, &term2 ); + sub128( rem1, 0, term1, term2, &rem1, &rem2 ); + while ( (sbits64) rem1 < 0 ) { + --zSig1; + add128( rem1, rem2, 0, bSig, &rem1, &rem2 ); + } + zSig1 |= ( ( rem1 | rem2 ) != 0 ); + } + return + roundAndPackFloatx80( + floatx80_rounding_precision, zSign, zExp, zSig0, zSig1 ); } @@ -3485,92 +3485,92 @@ floatx80 floatx80_div( floatx80 a, floatx80 b ) floatx80 floatx80_rem( floatx80 a, floatx80 b ) { - flag aSign, zSign; - int32 aExp, bExp, expDiff; - bits64 aSig0, aSig1, bSig; - bits64 q, term0, term1, alternateASig0, alternateASig1; - floatx80 z; - - aSig0 = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - bSig = extractFloatx80Frac( b ); - bExp = extractFloatx80Exp( b ); + flag aSign, zSign; + int32 aExp, bExp, expDiff; + bits64 aSig0, aSig1, bSig; + bits64 q, term0, term1, alternateASig0, alternateASig1; + floatx80 z; + + aSig0 = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + bSig = extractFloatx80Frac( b ); + bExp = extractFloatx80Exp( b ); // bSign = extractFloatx80Sign( b ); - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig0<<1 ) - || ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { - return propagateFloatx80NaN( a, b ); - } - goto invalid; - } - if ( bExp == 0x7FFF ) { - if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - if ( bSig == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = floatx80_default_nan_low; - z.high = floatx80_default_nan_high; - return z; - } - normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); - } - if ( aExp == 0 ) { - if ( (bits64) ( aSig0<<1 ) == 0 ) return a; - normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); - } - bSig |= LIT64( 0x8000000000000000 ); - zSign = aSign; - expDiff = aExp - bExp; - aSig1 = 0; - if ( expDiff < 0 ) { - if ( expDiff < -1 ) return a; - shift128Right( aSig0, 0, 1, &aSig0, &aSig1 ); - expDiff = 0; - } - q = ( bSig <= aSig0 ); - if ( q ) aSig0 -= bSig; - expDiff -= 64; - while ( 0 < expDiff ) { - q = estimateDiv128To64( aSig0, aSig1, bSig ); - q = ( 2 < q ) ? q - 2 : 0; - mul64To128( bSig, q, &term0, &term1 ); - sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); - shortShift128Left( aSig0, aSig1, 62, &aSig0, &aSig1 ); - expDiff -= 62; - } - expDiff += 64; - if ( 0 < expDiff ) { - q = estimateDiv128To64( aSig0, aSig1, bSig ); - q = ( 2 < q ) ? q - 2 : 0; - q >>= 64 - expDiff; - mul64To128( bSig, q<<( 64 - expDiff ), &term0, &term1 ); - sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); - shortShift128Left( 0, bSig, 64 - expDiff, &term0, &term1 ); - while ( le128( term0, term1, aSig0, aSig1 ) ) { - ++q; - sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); - } - } - else { - term1 = 0; - term0 = bSig; - } - sub128( term0, term1, aSig0, aSig1, &alternateASig0, &alternateASig1 ); - if ( lt128( alternateASig0, alternateASig1, aSig0, aSig1 ) - || ( eq128( alternateASig0, alternateASig1, aSig0, aSig1 ) - && ( q & 1 ) ) - ) { - aSig0 = alternateASig0; - aSig1 = alternateASig1; - zSign = ! zSign; - } - return - normalizeRoundAndPackFloatx80( - 80, zSign, bExp + expDiff, aSig0, aSig1 ); + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig0<<1 ) + || ( ( bExp == 0x7FFF ) && (bits64) ( bSig<<1 ) ) ) { + return propagateFloatx80NaN( a, b ); + } + goto invalid; + } + if ( bExp == 0x7FFF ) { + if ( (bits64) ( bSig<<1 ) ) return propagateFloatx80NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + if ( bSig == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = floatx80_default_nan_low; + z.high = floatx80_default_nan_high; + return z; + } + normalizeFloatx80Subnormal( bSig, &bExp, &bSig ); + } + if ( aExp == 0 ) { + if ( (bits64) ( aSig0<<1 ) == 0 ) return a; + normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); + } + bSig |= LIT64( 0x8000000000000000 ); + zSign = aSign; + expDiff = aExp - bExp; + aSig1 = 0; + if ( expDiff < 0 ) { + if ( expDiff < -1 ) return a; + shift128Right( aSig0, 0, 1, &aSig0, &aSig1 ); + expDiff = 0; + } + q = ( bSig <= aSig0 ); + if ( q ) aSig0 -= bSig; + expDiff -= 64; + while ( 0 < expDiff ) { + q = estimateDiv128To64( aSig0, aSig1, bSig ); + q = ( 2 < q ) ? q - 2 : 0; + mul64To128( bSig, q, &term0, &term1 ); + sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); + shortShift128Left( aSig0, aSig1, 62, &aSig0, &aSig1 ); + expDiff -= 62; + } + expDiff += 64; + if ( 0 < expDiff ) { + q = estimateDiv128To64( aSig0, aSig1, bSig ); + q = ( 2 < q ) ? q - 2 : 0; + q >>= 64 - expDiff; + mul64To128( bSig, q<<( 64 - expDiff ), &term0, &term1 ); + sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); + shortShift128Left( 0, bSig, 64 - expDiff, &term0, &term1 ); + while ( le128( term0, term1, aSig0, aSig1 ) ) { + ++q; + sub128( aSig0, aSig1, term0, term1, &aSig0, &aSig1 ); + } + } + else { + term1 = 0; + term0 = bSig; + } + sub128( term0, term1, aSig0, aSig1, &alternateASig0, &alternateASig1 ); + if ( lt128( alternateASig0, alternateASig1, aSig0, aSig1 ) + || ( eq128( alternateASig0, alternateASig1, aSig0, aSig1 ) + && ( q & 1 ) ) + ) { + aSig0 = alternateASig0; + aSig1 = alternateASig1; + zSign = ! zSign; + } + return + normalizeRoundAndPackFloatx80( + 80, zSign, bExp + expDiff, aSig0, aSig1 ); } @@ -3582,65 +3582,65 @@ floatx80 floatx80_rem( floatx80 a, floatx80 b ) floatx80 floatx80_sqrt( floatx80 a ) { - flag aSign; - int32 aExp, zExp; - bits64 aSig0, aSig1, zSig0, zSig1, doubleZSig0; - bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; - floatx80 z; - - aSig0 = extractFloatx80Frac( a ); - aExp = extractFloatx80Exp( a ); - aSign = extractFloatx80Sign( a ); - if ( aExp == 0x7FFF ) { - if ( (bits64) ( aSig0<<1 ) ) return propagateFloatx80NaN( a, a ); - if ( ! aSign ) return a; - goto invalid; - } - if ( aSign ) { - if ( ( aExp | aSig0 ) == 0 ) return a; - invalid: - float_raise( float_flag_invalid ); - z.low = floatx80_default_nan_low; - z.high = floatx80_default_nan_high; - return z; - } - if ( aExp == 0 ) { - if ( aSig0 == 0 ) return packFloatx80( 0, 0, 0 ); - normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); - } - zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFF; - zSig0 = estimateSqrt32( aExp, aSig0>>32 ); - shift128Right( aSig0, 0, 2 + ( aExp & 1 ), &aSig0, &aSig1 ); - zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); - doubleZSig0 = zSig0<<1; - mul64To128( zSig0, zSig0, &term0, &term1 ); - sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); - while ( (sbits64) rem0 < 0 ) { - --zSig0; - doubleZSig0 -= 2; - add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); - } - zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 ); - if ( ( zSig1 & LIT64( 0x3FFFFFFFFFFFFFFF ) ) <= 5 ) { - if ( zSig1 == 0 ) zSig1 = 1; - mul64To128( doubleZSig0, zSig1, &term1, &term2 ); - sub128( rem1, 0, term1, term2, &rem1, &rem2 ); - mul64To128( zSig1, zSig1, &term2, &term3 ); - sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); - while ( (sbits64) rem1 < 0 ) { - --zSig1; - shortShift128Left( 0, zSig1, 1, &term2, &term3 ); - term3 |= 1; - term2 |= doubleZSig0; - add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 ); - } - zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); - } - shortShift128Left( 0, zSig1, 1, &zSig0, &zSig1 ); - zSig0 |= doubleZSig0; - return - roundAndPackFloatx80( - floatx80_rounding_precision, 0, zExp, zSig0, zSig1 ); + flag aSign; + int32 aExp, zExp; + bits64 aSig0, aSig1, zSig0, zSig1, doubleZSig0; + bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; + floatx80 z; + + aSig0 = extractFloatx80Frac( a ); + aExp = extractFloatx80Exp( a ); + aSign = extractFloatx80Sign( a ); + if ( aExp == 0x7FFF ) { + if ( (bits64) ( aSig0<<1 ) ) return propagateFloatx80NaN( a, a ); + if ( ! aSign ) return a; + goto invalid; + } + if ( aSign ) { + if ( ( aExp | aSig0 ) == 0 ) return a; + invalid: + float_raise( float_flag_invalid ); + z.low = floatx80_default_nan_low; + z.high = floatx80_default_nan_high; + return z; + } + if ( aExp == 0 ) { + if ( aSig0 == 0 ) return packFloatx80( 0, 0, 0 ); + normalizeFloatx80Subnormal( aSig0, &aExp, &aSig0 ); + } + zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFF; + zSig0 = estimateSqrt32( aExp, aSig0>>32 ); + shift128Right( aSig0, 0, 2 + ( aExp & 1 ), &aSig0, &aSig1 ); + zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); + doubleZSig0 = zSig0<<1; + mul64To128( zSig0, zSig0, &term0, &term1 ); + sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); + while ( (sbits64) rem0 < 0 ) { + --zSig0; + doubleZSig0 -= 2; + add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); + } + zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 ); + if ( ( zSig1 & LIT64( 0x3FFFFFFFFFFFFFFF ) ) <= 5 ) { + if ( zSig1 == 0 ) zSig1 = 1; + mul64To128( doubleZSig0, zSig1, &term1, &term2 ); + sub128( rem1, 0, term1, term2, &rem1, &rem2 ); + mul64To128( zSig1, zSig1, &term2, &term3 ); + sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); + while ( (sbits64) rem1 < 0 ) { + --zSig1; + shortShift128Left( 0, zSig1, 1, &term2, &term3 ); + term3 |= 1; + term2 |= doubleZSig0; + add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 ); + } + zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); + } + shortShift128Left( 0, zSig1, 1, &zSig0, &zSig1 ); + zSig0 |= doubleZSig0; + return + roundAndPackFloatx80( + floatx80_rounding_precision, 0, zExp, zSig0, zSig1 ); } @@ -3654,23 +3654,23 @@ floatx80 floatx80_sqrt( floatx80 a ) flag floatx80_eq( floatx80 a, floatx80 b ) { - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - if ( floatx80_is_signaling_nan( a ) - || floatx80_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - return - ( a.low == b.low ) - && ( ( a.high == b.high ) - || ( ( a.low == 0 ) - && ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) - ); + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + if ( floatx80_is_signaling_nan( a ) + || floatx80_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + return + ( a.low == b.low ) + && ( ( a.high == b.high ) + || ( ( a.low == 0 ) + && ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) + ); } @@ -3683,27 +3683,27 @@ flag floatx80_eq( floatx80 a, floatx80 b ) flag floatx80_le( floatx80 a, floatx80 b ) { - flag aSign, bSign; - - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign != bSign ) { - return - aSign - || ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - == 0 ); - } - return - aSign ? le128( b.high, b.low, a.high, a.low ) - : le128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign != bSign ) { + return + aSign + || ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + == 0 ); + } + return + aSign ? le128( b.high, b.low, a.high, a.low ) + : le128( a.high, a.low, b.high, b.low ); } @@ -3716,27 +3716,27 @@ flag floatx80_le( floatx80 a, floatx80 b ) flag floatx80_lt( floatx80 a, floatx80 b ) { - flag aSign, bSign; - - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign != bSign ) { - return - aSign - && ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - != 0 ); - } - return - aSign ? lt128( b.high, b.low, a.high, a.low ) - : lt128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign != bSign ) { + return + aSign + && ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + != 0 ); + } + return + aSign ? lt128( b.high, b.low, a.high, a.low ) + : lt128( a.high, a.low, b.high, b.low ); } @@ -3750,20 +3750,20 @@ flag floatx80_lt( floatx80 a, floatx80 b ) flag floatx80_eq_signaling( floatx80 a, floatx80 b ) { - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - return - ( a.low == b.low ) - && ( ( a.high == b.high ) - || ( ( a.low == 0 ) - && ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) - ); + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + return + ( a.low == b.low ) + && ( ( a.high == b.high ) + || ( ( a.low == 0 ) + && ( (bits16) ( ( a.high | b.high )<<1 ) == 0 ) ) + ); } @@ -3776,30 +3776,30 @@ flag floatx80_eq_signaling( floatx80 a, floatx80 b ) flag floatx80_le_quiet( floatx80 a, floatx80 b ) { - flag aSign, bSign; - - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - if ( floatx80_is_signaling_nan( a ) - || floatx80_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign != bSign ) { - return - aSign - || ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - == 0 ); - } - return - aSign ? le128( b.high, b.low, a.high, a.low ) - : le128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + if ( floatx80_is_signaling_nan( a ) + || floatx80_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign != bSign ) { + return + aSign + || ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + == 0 ); + } + return + aSign ? le128( b.high, b.low, a.high, a.low ) + : le128( a.high, a.low, b.high, b.low ); } @@ -3812,30 +3812,30 @@ flag floatx80_le_quiet( floatx80 a, floatx80 b ) flag floatx80_lt_quiet( floatx80 a, floatx80 b ) { - flag aSign, bSign; - - if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( a )<<1 ) ) - || ( ( extractFloatx80Exp( b ) == 0x7FFF ) - && (bits64) ( extractFloatx80Frac( b )<<1 ) ) - ) { - if ( floatx80_is_signaling_nan( a ) - || floatx80_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloatx80Sign( a ); - bSign = extractFloatx80Sign( b ); - if ( aSign != bSign ) { - return - aSign - && ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - != 0 ); - } - return - aSign ? lt128( b.high, b.low, a.high, a.low ) - : lt128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloatx80Exp( a ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( a )<<1 ) ) + || ( ( extractFloatx80Exp( b ) == 0x7FFF ) + && (bits64) ( extractFloatx80Frac( b )<<1 ) ) + ) { + if ( floatx80_is_signaling_nan( a ) + || floatx80_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloatx80Sign( a ); + bSign = extractFloatx80Sign( b ); + if ( aSign != bSign ) { + return + aSign + && ( ( ( (bits16) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + != 0 ); + } + return + aSign ? lt128( b.high, b.low, a.high, a.low ) + : lt128( a.high, a.low, b.high, b.low ); } @@ -3855,20 +3855,20 @@ flag floatx80_lt_quiet( floatx80 a, floatx80 b ) int32 float128_to_int32( float128 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig0, aSig1; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) aSign = 0; - if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); - aSig0 |= ( aSig1 != 0 ); - shiftCount = 0x4028 - aExp; - if ( 0 < shiftCount ) shift64RightJamming( aSig0, shiftCount, &aSig0 ); - return roundAndPackInt32( aSign, aSig0 ); + flag aSign; + int32 aExp, shiftCount; + bits64 aSig0, aSig1; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) aSign = 0; + if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); + aSig0 |= ( aSig1 != 0 ); + shiftCount = 0x4028 - aExp; + if ( 0 < shiftCount ) shift64RightJamming( aSig0, shiftCount, &aSig0 ); + return roundAndPackInt32( aSign, aSig0 ); } @@ -3884,39 +3884,39 @@ int32 float128_to_int32( float128 a ) int32 float128_to_int32_round_to_zero( float128 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig0, aSig1, savedASig; - int32 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - aSig0 |= ( aSig1 != 0 ); - if ( 0x401E < aExp ) { - if ( ( aExp == 0x7FFF ) && aSig0 ) aSign = 0; - goto invalid; - } - else if ( aExp < 0x3FFF ) { - if ( aExp || aSig0 ) float_exception_flags |= float_flag_inexact; - return 0; - } - aSig0 |= LIT64( 0x0001000000000000 ); - shiftCount = 0x402F - aExp; - savedASig = aSig0; - aSig0 >>= shiftCount; - z = aSig0; - if ( aSign ) z = - z; - if ( ( z < 0 ) ^ aSign ) { - invalid: - float_raise( float_flag_invalid ); - return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; - } - if ( ( aSig0<<shiftCount ) != savedASig ) { - float_exception_flags |= float_flag_inexact; - } - return z; + flag aSign; + int32 aExp, shiftCount; + bits64 aSig0, aSig1, savedASig; + int32 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + aSig0 |= ( aSig1 != 0 ); + if ( 0x401E < aExp ) { + if ( ( aExp == 0x7FFF ) && aSig0 ) aSign = 0; + goto invalid; + } + else if ( aExp < 0x3FFF ) { + if ( aExp || aSig0 ) float_exception_flags |= float_flag_inexact; + return 0; + } + aSig0 |= LIT64( 0x0001000000000000 ); + shiftCount = 0x402F - aExp; + savedASig = aSig0; + aSig0 >>= shiftCount; + z = aSig0; + if ( aSign ) z = - z; + if ( ( z < 0 ) ^ aSign ) { + invalid: + float_raise( float_flag_invalid ); + return aSign ? (sbits32) 0x80000000 : 0x7FFFFFFF; + } + if ( ( aSig0<<shiftCount ) != savedASig ) { + float_exception_flags |= float_flag_inexact; + } + return z; } @@ -3932,34 +3932,34 @@ int32 float128_to_int32_round_to_zero( float128 a ) int64 float128_to_int64( float128 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig0, aSig1; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); - shiftCount = 0x402F - aExp; - if ( shiftCount <= 0 ) { - if ( 0x403E < aExp ) { - float_raise( float_flag_invalid ); - if ( ! aSign - || ( ( aExp == 0x7FFF ) - && ( aSig1 || ( aSig0 != LIT64( 0x0001000000000000 ) ) ) - ) - ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - shortShift128Left( aSig0, aSig1, - shiftCount, &aSig0, &aSig1 ); - } - else { - shift64ExtraRightJamming( aSig0, aSig1, shiftCount, &aSig0, &aSig1 ); - } - return roundAndPackInt64( aSign, aSig0, aSig1 ); + flag aSign; + int32 aExp, shiftCount; + bits64 aSig0, aSig1; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); + shiftCount = 0x402F - aExp; + if ( shiftCount <= 0 ) { + if ( 0x403E < aExp ) { + float_raise( float_flag_invalid ); + if ( ! aSign + || ( ( aExp == 0x7FFF ) + && ( aSig1 || ( aSig0 != LIT64( 0x0001000000000000 ) ) ) + ) + ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + shortShift128Left( aSig0, aSig1, - shiftCount, &aSig0, &aSig1 ); + } + else { + shift64ExtraRightJamming( aSig0, aSig1, shiftCount, &aSig0, &aSig1 ); + } + return roundAndPackInt64( aSign, aSig0, aSig1 ); } @@ -3975,52 +3975,52 @@ int64 float128_to_int64( float128 a ) int64 float128_to_int64_round_to_zero( float128 a ) { - flag aSign; - int32 aExp, shiftCount; - bits64 aSig0, aSig1; - int64 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); - shiftCount = aExp - 0x402F; - if ( 0 < shiftCount ) { - if ( 0x403E <= aExp ) { - aSig0 &= LIT64( 0x0000FFFFFFFFFFFF ); - if ( ( a.high == LIT64( 0xC03E000000000000 ) ) - && ( aSig1 < LIT64( 0x0002000000000000 ) ) ) { - if ( aSig1 ) float_exception_flags |= float_flag_inexact; - } - else { - float_raise( float_flag_invalid ); - if ( ! aSign || ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) ) { - return LIT64( 0x7FFFFFFFFFFFFFFF ); - } - } - return (sbits64) LIT64( 0x8000000000000000 ); - } - z = ( aSig0<<shiftCount ) | ( aSig1>>( ( - shiftCount ) & 63 ) ); - if ( (bits64) ( aSig1<<shiftCount ) ) { - float_exception_flags |= float_flag_inexact; - } - } - else { - if ( aExp < 0x3FFF ) { - if ( aExp | aSig0 | aSig1 ) { - float_exception_flags |= float_flag_inexact; - } - return 0; - } - z = aSig0>>( - shiftCount ); - if ( aSig1 - || ( shiftCount && (bits64) ( aSig0<<( shiftCount & 63 ) ) ) ) { - float_exception_flags |= float_flag_inexact; - } - } - if ( aSign ) z = - z; - return z; + flag aSign; + int32 aExp, shiftCount; + bits64 aSig0, aSig1; + int64 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp ) aSig0 |= LIT64( 0x0001000000000000 ); + shiftCount = aExp - 0x402F; + if ( 0 < shiftCount ) { + if ( 0x403E <= aExp ) { + aSig0 &= LIT64( 0x0000FFFFFFFFFFFF ); + if ( ( a.high == LIT64( 0xC03E000000000000 ) ) + && ( aSig1 < LIT64( 0x0002000000000000 ) ) ) { + if ( aSig1 ) float_exception_flags |= float_flag_inexact; + } + else { + float_raise( float_flag_invalid ); + if ( ! aSign || ( ( aExp == 0x7FFF ) && ( aSig0 | aSig1 ) ) ) { + return LIT64( 0x7FFFFFFFFFFFFFFF ); + } + } + return (sbits64) LIT64( 0x8000000000000000 ); + } + z = ( aSig0<<shiftCount ) | ( aSig1>>( ( - shiftCount ) & 63 ) ); + if ( (bits64) ( aSig1<<shiftCount ) ) { + float_exception_flags |= float_flag_inexact; + } + } + else { + if ( aExp < 0x3FFF ) { + if ( aExp | aSig0 | aSig1 ) { + float_exception_flags |= float_flag_inexact; + } + return 0; + } + z = aSig0>>( - shiftCount ); + if ( aSig1 + || ( shiftCount && (bits64) ( aSig0<<( shiftCount & 63 ) ) ) ) { + float_exception_flags |= float_flag_inexact; + } + } + if ( aSign ) z = - z; + return z; } @@ -4033,29 +4033,29 @@ int64 float128_to_int64_round_to_zero( float128 a ) float32 float128_to_float32( float128 a ) { - flag aSign; - int32 aExp; - bits64 aSig0, aSig1; - bits32 zSig; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) { - return commonNaNToFloat32( float128ToCommonNaN( a ) ); - } - return packFloat32( aSign, 0xFF, 0 ); - } - aSig0 |= ( aSig1 != 0 ); - shift64RightJamming( aSig0, 18, &aSig0 ); - zSig = aSig0; - if ( aExp || zSig ) { - zSig |= 0x40000000; - aExp -= 0x3F81; - } - return roundAndPackFloat32( aSign, aExp, zSig ); + flag aSign; + int32 aExp; + bits64 aSig0, aSig1; + bits32 zSig; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) { + return commonNaNToFloat32( float128ToCommonNaN( a ) ); + } + return packFloat32( aSign, 0xFF, 0 ); + } + aSig0 |= ( aSig1 != 0 ); + shift64RightJamming( aSig0, 18, &aSig0 ); + zSig = aSig0; + if ( aExp || zSig ) { + zSig |= 0x40000000; + aExp -= 0x3F81; + } + return roundAndPackFloat32( aSign, aExp, zSig ); } @@ -4068,27 +4068,27 @@ float32 float128_to_float32( float128 a ) float64 float128_to_float64( float128 a ) { - flag aSign; - int32 aExp; - bits64 aSig0, aSig1; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) { - return commonNaNToFloat64( float128ToCommonNaN( a ) ); - } - return packFloat64( aSign, 0x7FF, 0 ); - } - shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 ); - aSig0 |= ( aSig1 != 0 ); - if ( aExp || aSig0 ) { - aSig0 |= LIT64( 0x4000000000000000 ); - aExp -= 0x3C01; - } - return roundAndPackFloat64( aSign, aExp, aSig0 ); + flag aSign; + int32 aExp; + bits64 aSig0, aSig1; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) { + return commonNaNToFloat64( float128ToCommonNaN( a ) ); + } + return packFloat64( aSign, 0x7FF, 0 ); + } + shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 ); + aSig0 |= ( aSig1 != 0 ); + if ( aExp || aSig0 ) { + aSig0 |= LIT64( 0x4000000000000000 ); + aExp -= 0x3C01; + } + return roundAndPackFloat64( aSign, aExp, aSig0 ); } @@ -4103,29 +4103,29 @@ float64 float128_to_float64( float128 a ) floatx80 float128_to_floatx80( float128 a ) { - flag aSign; - int32 aExp; - bits64 aSig0, aSig1; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) { - return commonNaNToFloatx80( float128ToCommonNaN( a ) ); - } - return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); - } - if ( aExp == 0 ) { - if ( ( aSig0 | aSig1 ) == 0 ) return packFloatx80( aSign, 0, 0 ); - normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); - } - else { - aSig0 |= LIT64( 0x0001000000000000 ); - } - shortShift128Left( aSig0, aSig1, 15, &aSig0, &aSig1 ); - return roundAndPackFloatx80( 80, aSign, aExp, aSig0, aSig1 ); + flag aSign; + int32 aExp; + bits64 aSig0, aSig1; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) { + return commonNaNToFloatx80( float128ToCommonNaN( a ) ); + } + return packFloatx80( aSign, 0x7FFF, LIT64( 0x8000000000000000 ) ); + } + if ( aExp == 0 ) { + if ( ( aSig0 | aSig1 ) == 0 ) return packFloatx80( aSign, 0, 0 ); + normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); + } + else { + aSig0 |= LIT64( 0x0001000000000000 ); + } + shortShift128Left( aSig0, aSig1, 15, &aSig0, &aSig1 ); + return roundAndPackFloatx80( 80, aSign, aExp, aSig0, aSig1 ); } @@ -4140,97 +4140,97 @@ floatx80 float128_to_floatx80( float128 a ) float128 float128_round_to_int( float128 a ) { - flag aSign; - int32 aExp; - bits64 lastBitMask, roundBitsMask; - int8 roundingMode; - float128 z; - - aExp = extractFloat128Exp( a ); - if ( 0x402F <= aExp ) { - if ( 0x406F <= aExp ) { - if ( ( aExp == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) - ) { - return propagateFloat128NaN( a, a ); - } - return a; - } - lastBitMask = 1; - lastBitMask = ( lastBitMask<<( 0x406E - aExp ) )<<1; - roundBitsMask = lastBitMask - 1; - z = a; - roundingMode = float_rounding_mode; - if ( roundingMode == float_round_nearest_even ) { - if ( lastBitMask ) { - add128( z.high, z.low, 0, lastBitMask>>1, &z.high, &z.low ); - if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; - } - else { - if ( (sbits64) z.low < 0 ) { - ++z.high; - if ( (bits64) ( z.low<<1 ) == 0 ) z.high &= ~1; - } - } - } - else if ( roundingMode != float_round_to_zero ) { - if ( extractFloat128Sign( z ) - ^ ( roundingMode == float_round_up ) ) { - add128( z.high, z.low, 0, roundBitsMask, &z.high, &z.low ); - } - } - z.low &= ~ roundBitsMask; - } - else { - if ( aExp < 0x3FFF ) { - if ( ( ( (bits64) ( a.high<<1 ) ) | a.low ) == 0 ) return a; - float_exception_flags |= float_flag_inexact; - aSign = extractFloat128Sign( a ); - switch ( float_rounding_mode ) { - case float_round_nearest_even: - if ( ( aExp == 0x3FFE ) - && ( extractFloat128Frac0( a ) - | extractFloat128Frac1( a ) ) - ) { - return packFloat128( aSign, 0x3FFF, 0, 0 ); - } - break; - case float_round_down: - return - aSign ? packFloat128( 1, 0x3FFF, 0, 0 ) - : packFloat128( 0, 0, 0, 0 ); - case float_round_up: - return - aSign ? packFloat128( 1, 0, 0, 0 ) - : packFloat128( 0, 0x3FFF, 0, 0 ); - } - return packFloat128( aSign, 0, 0, 0 ); - } - lastBitMask = 1; - lastBitMask <<= 0x402F - aExp; - roundBitsMask = lastBitMask - 1; - z.low = 0; - z.high = a.high; - roundingMode = float_rounding_mode; - if ( roundingMode == float_round_nearest_even ) { - z.high += lastBitMask>>1; - if ( ( ( z.high & roundBitsMask ) | a.low ) == 0 ) { - z.high &= ~ lastBitMask; - } - } - else if ( roundingMode != float_round_to_zero ) { - if ( extractFloat128Sign( z ) - ^ ( roundingMode == float_round_up ) ) { - z.high |= ( a.low != 0 ); - z.high += roundBitsMask; - } - } - z.high &= ~ roundBitsMask; - } - if ( ( z.low != a.low ) || ( z.high != a.high ) ) { - float_exception_flags |= float_flag_inexact; - } - return z; + flag aSign; + int32 aExp; + bits64 lastBitMask, roundBitsMask; + int8 roundingMode; + float128 z; + + aExp = extractFloat128Exp( a ); + if ( 0x402F <= aExp ) { + if ( 0x406F <= aExp ) { + if ( ( aExp == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) + ) { + return propagateFloat128NaN( a, a ); + } + return a; + } + lastBitMask = 1; + lastBitMask = ( lastBitMask<<( 0x406E - aExp ) )<<1; + roundBitsMask = lastBitMask - 1; + z = a; + roundingMode = float_rounding_mode; + if ( roundingMode == float_round_nearest_even ) { + if ( lastBitMask ) { + add128( z.high, z.low, 0, lastBitMask>>1, &z.high, &z.low ); + if ( ( z.low & roundBitsMask ) == 0 ) z.low &= ~ lastBitMask; + } + else { + if ( (sbits64) z.low < 0 ) { + ++z.high; + if ( (bits64) ( z.low<<1 ) == 0 ) z.high &= ~1; + } + } + } + else if ( roundingMode != float_round_to_zero ) { + if ( extractFloat128Sign( z ) + ^ ( roundingMode == float_round_up ) ) { + add128( z.high, z.low, 0, roundBitsMask, &z.high, &z.low ); + } + } + z.low &= ~ roundBitsMask; + } + else { + if ( aExp < 0x3FFF ) { + if ( ( ( (bits64) ( a.high<<1 ) ) | a.low ) == 0 ) return a; + float_exception_flags |= float_flag_inexact; + aSign = extractFloat128Sign( a ); + switch ( float_rounding_mode ) { + case float_round_nearest_even: + if ( ( aExp == 0x3FFE ) + && ( extractFloat128Frac0( a ) + | extractFloat128Frac1( a ) ) + ) { + return packFloat128( aSign, 0x3FFF, 0, 0 ); + } + break; + case float_round_down: + return + aSign ? packFloat128( 1, 0x3FFF, 0, 0 ) + : packFloat128( 0, 0, 0, 0 ); + case float_round_up: + return + aSign ? packFloat128( 1, 0, 0, 0 ) + : packFloat128( 0, 0x3FFF, 0, 0 ); + } + return packFloat128( aSign, 0, 0, 0 ); + } + lastBitMask = 1; + lastBitMask <<= 0x402F - aExp; + roundBitsMask = lastBitMask - 1; + z.low = 0; + z.high = a.high; + roundingMode = float_rounding_mode; + if ( roundingMode == float_round_nearest_even ) { + z.high += lastBitMask>>1; + if ( ( ( z.high & roundBitsMask ) | a.low ) == 0 ) { + z.high &= ~ lastBitMask; + } + } + else if ( roundingMode != float_round_to_zero ) { + if ( extractFloat128Sign( z ) + ^ ( roundingMode == float_round_up ) ) { + z.high |= ( a.low != 0 ); + z.high += roundBitsMask; + } + } + z.high &= ~ roundBitsMask; + } + if ( ( z.low != a.low ) || ( z.high != a.high ) ) { + float_exception_flags |= float_flag_inexact; + } + return z; } @@ -4244,71 +4244,71 @@ float128 float128_round_to_int( float128 a ) static float128 addFloat128Sigs( float128 a, float128 b, flag zSign ) { - int32 aExp, bExp, zExp; - bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; - int32 expDiff; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - bSig1 = extractFloat128Frac1( b ); - bSig0 = extractFloat128Frac0( b ); - bExp = extractFloat128Exp( b ); - expDiff = aExp - bExp; - if ( 0 < expDiff ) { - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig0 |= LIT64( 0x0001000000000000 ); - } - shift128ExtraRightJamming( - bSig0, bSig1, 0, expDiff, &bSig0, &bSig1, &zSig2 ); - zExp = aExp; - } - else if ( expDiff < 0 ) { - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig0 |= LIT64( 0x0001000000000000 ); - } - shift128ExtraRightJamming( - aSig0, aSig1, 0, - expDiff, &aSig0, &aSig1, &zSig2 ); - zExp = bExp; - } - else { - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 | bSig0 | bSig1 ) { - return propagateFloat128NaN( a, b ); - } - return a; - } - add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); - if ( aExp == 0 ) return packFloat128( zSign, 0, zSig0, zSig1 ); - zSig2 = 0; - zSig0 |= LIT64( 0x0002000000000000 ); - zExp = aExp; - goto shiftRight1; - } - aSig0 |= LIT64( 0x0001000000000000 ); - add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); - --zExp; - if ( zSig0 < LIT64( 0x0002000000000000 ) ) goto roundAndPack; - ++zExp; - shiftRight1: - shift128ExtraRightJamming( - zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 ); - roundAndPack: - return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); + int32 aExp, bExp, zExp; + bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; + int32 expDiff; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + bSig1 = extractFloat128Frac1( b ); + bSig0 = extractFloat128Frac0( b ); + bExp = extractFloat128Exp( b ); + expDiff = aExp - bExp; + if ( 0 < expDiff ) { + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig0 |= LIT64( 0x0001000000000000 ); + } + shift128ExtraRightJamming( + bSig0, bSig1, 0, expDiff, &bSig0, &bSig1, &zSig2 ); + zExp = aExp; + } + else if ( expDiff < 0 ) { + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig0 |= LIT64( 0x0001000000000000 ); + } + shift128ExtraRightJamming( + aSig0, aSig1, 0, - expDiff, &aSig0, &aSig1, &zSig2 ); + zExp = bExp; + } + else { + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 | bSig0 | bSig1 ) { + return propagateFloat128NaN( a, b ); + } + return a; + } + add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); + if ( aExp == 0 ) return packFloat128( zSign, 0, zSig0, zSig1 ); + zSig2 = 0; + zSig0 |= LIT64( 0x0002000000000000 ); + zExp = aExp; + goto shiftRight1; + } + aSig0 |= LIT64( 0x0001000000000000 ); + add128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); + --zExp; + if ( zSig0 < LIT64( 0x0002000000000000 ) ) goto roundAndPack; + ++zExp; + shiftRight1: + shift128ExtraRightJamming( + zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 ); + roundAndPack: + return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); } @@ -4322,77 +4322,77 @@ static float128 addFloat128Sigs( float128 a, float128 b, flag zSign ) static float128 subFloat128Sigs( float128 a, float128 b, flag zSign ) { - int32 aExp, bExp, zExp; - bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1; - int32 expDiff; - float128 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - bSig1 = extractFloat128Frac1( b ); - bSig0 = extractFloat128Frac0( b ); - bExp = extractFloat128Exp( b ); - expDiff = aExp - bExp; - shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 ); - shortShift128Left( bSig0, bSig1, 14, &bSig0, &bSig1 ); - if ( 0 < expDiff ) goto aExpBigger; - if ( expDiff < 0 ) goto bExpBigger; - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 | bSig0 | bSig1 ) { - return propagateFloat128NaN( a, b ); - } - float_raise( float_flag_invalid ); - z.low = float128_default_nan_low; - z.high = float128_default_nan_high; - return z; - } - if ( aExp == 0 ) { - aExp = 1; - bExp = 1; - } - if ( bSig0 < aSig0 ) goto aBigger; - if ( aSig0 < bSig0 ) goto bBigger; - if ( bSig1 < aSig1 ) goto aBigger; - if ( aSig1 < bSig1 ) goto bBigger; - return packFloat128( float_rounding_mode == float_round_down, 0, 0, 0 ); - bExpBigger: - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - return packFloat128( zSign ^ 1, 0x7FFF, 0, 0 ); - } - if ( aExp == 0 ) { - ++expDiff; - } - else { - aSig0 |= LIT64( 0x4000000000000000 ); - } - shift128RightJamming( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); - bSig0 |= LIT64( 0x4000000000000000 ); - bBigger: - sub128( bSig0, bSig1, aSig0, aSig1, &zSig0, &zSig1 ); - zExp = bExp; - zSign ^= 1; - goto normalizeRoundAndPack; - aExpBigger: - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - --expDiff; - } - else { - bSig0 |= LIT64( 0x4000000000000000 ); - } - shift128RightJamming( bSig0, bSig1, expDiff, &bSig0, &bSig1 ); - aSig0 |= LIT64( 0x4000000000000000 ); - aBigger: - sub128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); - zExp = aExp; - normalizeRoundAndPack: - --zExp; - return normalizeRoundAndPackFloat128( zSign, zExp - 14, zSig0, zSig1 ); + int32 aExp, bExp, zExp; + bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1; + int32 expDiff; + float128 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + bSig1 = extractFloat128Frac1( b ); + bSig0 = extractFloat128Frac0( b ); + bExp = extractFloat128Exp( b ); + expDiff = aExp - bExp; + shortShift128Left( aSig0, aSig1, 14, &aSig0, &aSig1 ); + shortShift128Left( bSig0, bSig1, 14, &bSig0, &bSig1 ); + if ( 0 < expDiff ) goto aExpBigger; + if ( expDiff < 0 ) goto bExpBigger; + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 | bSig0 | bSig1 ) { + return propagateFloat128NaN( a, b ); + } + float_raise( float_flag_invalid ); + z.low = float128_default_nan_low; + z.high = float128_default_nan_high; + return z; + } + if ( aExp == 0 ) { + aExp = 1; + bExp = 1; + } + if ( bSig0 < aSig0 ) goto aBigger; + if ( aSig0 < bSig0 ) goto bBigger; + if ( bSig1 < aSig1 ) goto aBigger; + if ( aSig1 < bSig1 ) goto bBigger; + return packFloat128( float_rounding_mode == float_round_down, 0, 0, 0 ); + bExpBigger: + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + return packFloat128( zSign ^ 1, 0x7FFF, 0, 0 ); + } + if ( aExp == 0 ) { + ++expDiff; + } + else { + aSig0 |= LIT64( 0x4000000000000000 ); + } + shift128RightJamming( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); + bSig0 |= LIT64( 0x4000000000000000 ); + bBigger: + sub128( bSig0, bSig1, aSig0, aSig1, &zSig0, &zSig1 ); + zExp = bExp; + zSign ^= 1; + goto normalizeRoundAndPack; + aExpBigger: + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + --expDiff; + } + else { + bSig0 |= LIT64( 0x4000000000000000 ); + } + shift128RightJamming( bSig0, bSig1, expDiff, &bSig0, &bSig1 ); + aSig0 |= LIT64( 0x4000000000000000 ); + aBigger: + sub128( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1 ); + zExp = aExp; + normalizeRoundAndPack: + --zExp; + return normalizeRoundAndPackFloat128( zSign, zExp - 14, zSig0, zSig1 ); } @@ -4404,16 +4404,16 @@ static float128 subFloat128Sigs( float128 a, float128 b, flag zSign ) float128 float128_add( float128 a, float128 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign == bSign ) { - return addFloat128Sigs( a, b, aSign ); - } - else { - return subFloat128Sigs( a, b, aSign ); - } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign == bSign ) { + return addFloat128Sigs( a, b, aSign ); + } + else { + return subFloat128Sigs( a, b, aSign ); + } } @@ -4425,16 +4425,16 @@ float128 float128_add( float128 a, float128 b ) float128 float128_sub( float128 a, float128 b ) { - flag aSign, bSign; + flag aSign, bSign; - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign == bSign ) { - return subFloat128Sigs( a, b, aSign ); - } - else { - return addFloat128Sigs( a, b, aSign ); - } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign == bSign ) { + return subFloat128Sigs( a, b, aSign ); + } + else { + return addFloat128Sigs( a, b, aSign ); + } } @@ -4446,59 +4446,59 @@ float128 float128_sub( float128 a, float128 b ) float128 float128_mul( float128 a, float128 b ) { - flag aSign, bSign, zSign; - int32 aExp, bExp, zExp; - bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2, zSig3; - float128 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - bSig1 = extractFloat128Frac1( b ); - bSig0 = extractFloat128Frac0( b ); - bExp = extractFloat128Exp( b ); - bSign = extractFloat128Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FFF ) { - if ( ( aSig0 | aSig1 ) - || ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) { - return propagateFloat128NaN( a, b ); - } - if ( ( bExp | bSig0 | bSig1 ) == 0 ) goto invalid; - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - if ( ( aExp | aSig0 | aSig1 ) == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = float128_default_nan_low; - z.high = float128_default_nan_high; - return z; - } - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - if ( aExp == 0 ) { - if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); - normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); - } - if ( bExp == 0 ) { - if ( ( bSig0 | bSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); - normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); - } - zExp = aExp + bExp - 0x4000; - aSig0 |= LIT64( 0x0001000000000000 ); - shortShift128Left( bSig0, bSig1, 16, &bSig0, &bSig1 ); - mul128To256( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1, &zSig2, &zSig3 ); - add128( zSig0, zSig1, aSig0, aSig1, &zSig0, &zSig1 ); - zSig2 |= ( zSig3 != 0 ); - if ( LIT64( 0x0002000000000000 ) <= zSig0 ) { - shift128ExtraRightJamming( - zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 ); - ++zExp; - } - return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); + flag aSign, bSign, zSign; + int32 aExp, bExp, zExp; + bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2, zSig3; + float128 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + bSig1 = extractFloat128Frac1( b ); + bSig0 = extractFloat128Frac0( b ); + bExp = extractFloat128Exp( b ); + bSign = extractFloat128Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FFF ) { + if ( ( aSig0 | aSig1 ) + || ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) { + return propagateFloat128NaN( a, b ); + } + if ( ( bExp | bSig0 | bSig1 ) == 0 ) goto invalid; + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + if ( ( aExp | aSig0 | aSig1 ) == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = float128_default_nan_low; + z.high = float128_default_nan_high; + return z; + } + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + if ( aExp == 0 ) { + if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); + normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); + } + if ( bExp == 0 ) { + if ( ( bSig0 | bSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); + normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); + } + zExp = aExp + bExp - 0x4000; + aSig0 |= LIT64( 0x0001000000000000 ); + shortShift128Left( bSig0, bSig1, 16, &bSig0, &bSig1 ); + mul128To256( aSig0, aSig1, bSig0, bSig1, &zSig0, &zSig1, &zSig2, &zSig3 ); + add128( zSig0, zSig1, aSig0, aSig1, &zSig0, &zSig1 ); + zSig2 |= ( zSig3 != 0 ); + if ( LIT64( 0x0002000000000000 ) <= zSig0 ) { + shift128ExtraRightJamming( + zSig0, zSig1, zSig2, 1, &zSig0, &zSig1, &zSig2 ); + ++zExp; + } + return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); } @@ -4510,79 +4510,79 @@ float128 float128_mul( float128 a, float128 b ) float128 float128_div( float128 a, float128 b ) { - flag aSign, bSign, zSign; - int32 aExp, bExp, zExp; - bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; - bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; - float128 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - bSig1 = extractFloat128Frac1( b ); - bSig0 = extractFloat128Frac0( b ); - bExp = extractFloat128Exp( b ); - bSign = extractFloat128Sign( b ); - zSign = aSign ^ bSign; - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - goto invalid; - } - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - return packFloat128( zSign, 0, 0, 0 ); - } - if ( bExp == 0 ) { - if ( ( bSig0 | bSig1 ) == 0 ) { - if ( ( aExp | aSig0 | aSig1 ) == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = float128_default_nan_low; - z.high = float128_default_nan_high; - return z; - } - float_raise( float_flag_divbyzero ); - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); - } - if ( aExp == 0 ) { - if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); - normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); - } - zExp = aExp - bExp + 0x3FFD; - shortShift128Left( - aSig0 | LIT64( 0x0001000000000000 ), aSig1, 15, &aSig0, &aSig1 ); - shortShift128Left( - bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); - if ( le128( bSig0, bSig1, aSig0, aSig1 ) ) { - shift128Right( aSig0, aSig1, 1, &aSig0, &aSig1 ); - ++zExp; - } - zSig0 = estimateDiv128To64( aSig0, aSig1, bSig0 ); - mul128By64To192( bSig0, bSig1, zSig0, &term0, &term1, &term2 ); - sub192( aSig0, aSig1, 0, term0, term1, term2, &rem0, &rem1, &rem2 ); - while ( (sbits64) rem0 < 0 ) { - --zSig0; - add192( rem0, rem1, rem2, 0, bSig0, bSig1, &rem0, &rem1, &rem2 ); - } - zSig1 = estimateDiv128To64( rem1, rem2, bSig0 ); - if ( ( zSig1 & 0x3FFF ) <= 4 ) { - mul128By64To192( bSig0, bSig1, zSig1, &term1, &term2, &term3 ); - sub192( rem1, rem2, 0, term1, term2, term3, &rem1, &rem2, &rem3 ); - while ( (sbits64) rem1 < 0 ) { - --zSig1; - add192( rem1, rem2, rem3, 0, bSig0, bSig1, &rem1, &rem2, &rem3 ); - } - zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); - } - shift128ExtraRightJamming( zSig0, zSig1, 0, 15, &zSig0, &zSig1, &zSig2 ); - return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); + flag aSign, bSign, zSign; + int32 aExp, bExp, zExp; + bits64 aSig0, aSig1, bSig0, bSig1, zSig0, zSig1, zSig2; + bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; + float128 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + bSig1 = extractFloat128Frac1( b ); + bSig0 = extractFloat128Frac0( b ); + bExp = extractFloat128Exp( b ); + bSign = extractFloat128Sign( b ); + zSign = aSign ^ bSign; + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, b ); + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + goto invalid; + } + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + return packFloat128( zSign, 0, 0, 0 ); + } + if ( bExp == 0 ) { + if ( ( bSig0 | bSig1 ) == 0 ) { + if ( ( aExp | aSig0 | aSig1 ) == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = float128_default_nan_low; + z.high = float128_default_nan_high; + return z; + } + float_raise( float_flag_divbyzero ); + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); + } + if ( aExp == 0 ) { + if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( zSign, 0, 0, 0 ); + normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); + } + zExp = aExp - bExp + 0x3FFD; + shortShift128Left( + aSig0 | LIT64( 0x0001000000000000 ), aSig1, 15, &aSig0, &aSig1 ); + shortShift128Left( + bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); + if ( le128( bSig0, bSig1, aSig0, aSig1 ) ) { + shift128Right( aSig0, aSig1, 1, &aSig0, &aSig1 ); + ++zExp; + } + zSig0 = estimateDiv128To64( aSig0, aSig1, bSig0 ); + mul128By64To192( bSig0, bSig1, zSig0, &term0, &term1, &term2 ); + sub192( aSig0, aSig1, 0, term0, term1, term2, &rem0, &rem1, &rem2 ); + while ( (sbits64) rem0 < 0 ) { + --zSig0; + add192( rem0, rem1, rem2, 0, bSig0, bSig1, &rem0, &rem1, &rem2 ); + } + zSig1 = estimateDiv128To64( rem1, rem2, bSig0 ); + if ( ( zSig1 & 0x3FFF ) <= 4 ) { + mul128By64To192( bSig0, bSig1, zSig1, &term1, &term2, &term3 ); + sub192( rem1, rem2, 0, term1, term2, term3, &rem1, &rem2, &rem3 ); + while ( (sbits64) rem1 < 0 ) { + --zSig1; + add192( rem1, rem2, rem3, 0, bSig0, bSig1, &rem1, &rem2, &rem3 ); + } + zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); + } + shift128ExtraRightJamming( zSig0, zSig1, 0, 15, &zSig0, &zSig1, &zSig2 ); + return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); } @@ -4594,105 +4594,105 @@ float128 float128_div( float128 a, float128 b ) float128 float128_rem( float128 a, float128 b ) { - flag aSign, zSign; - int32 aExp, bExp, expDiff; - bits64 aSig0, aSig1, bSig0, bSig1, q, term0, term1, term2; - bits64 allZero, alternateASig0, alternateASig1, sigMean1; - sbits64 sigMean0; - float128 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - bSig1 = extractFloat128Frac1( b ); - bSig0 = extractFloat128Frac0( b ); - bExp = extractFloat128Exp( b ); + flag aSign, zSign; + int32 aExp, bExp, expDiff; + bits64 aSig0, aSig1, bSig0, bSig1, q, term0, term1, term2; + bits64 allZero, alternateASig0, alternateASig1, sigMean1; + sbits64 sigMean0; + float128 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + bSig1 = extractFloat128Frac1( b ); + bSig0 = extractFloat128Frac0( b ); + bExp = extractFloat128Exp( b ); // bSign = extractFloat128Sign( b ); - if ( aExp == 0x7FFF ) { - if ( ( aSig0 | aSig1 ) - || ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) { - return propagateFloat128NaN( a, b ); - } - goto invalid; - } - if ( bExp == 0x7FFF ) { - if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); - return a; - } - if ( bExp == 0 ) { - if ( ( bSig0 | bSig1 ) == 0 ) { - invalid: - float_raise( float_flag_invalid ); - z.low = float128_default_nan_low; - z.high = float128_default_nan_high; - return z; - } - normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); - } - if ( aExp == 0 ) { - if ( ( aSig0 | aSig1 ) == 0 ) return a; - normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); - } - expDiff = aExp - bExp; - if ( expDiff < -1 ) return a; - shortShift128Left( - aSig0 | LIT64( 0x0001000000000000 ), - aSig1, - 15 - ( expDiff < 0 ), - &aSig0, - &aSig1 - ); - shortShift128Left( - bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); - q = le128( bSig0, bSig1, aSig0, aSig1 ); - if ( q ) sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 ); - expDiff -= 64; - while ( 0 < expDiff ) { - q = estimateDiv128To64( aSig0, aSig1, bSig0 ); - q = ( 4 < q ) ? q - 4 : 0; - mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); - shortShift192Left( term0, term1, term2, 61, &term1, &term2, &allZero ); - shortShift128Left( aSig0, aSig1, 61, &aSig0, &allZero ); - sub128( aSig0, 0, term1, term2, &aSig0, &aSig1 ); - expDiff -= 61; - } - if ( -64 < expDiff ) { - q = estimateDiv128To64( aSig0, aSig1, bSig0 ); - q = ( 4 < q ) ? q - 4 : 0; - q >>= - expDiff; - shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 ); - expDiff += 52; - if ( expDiff < 0 ) { - shift128Right( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); - } - else { - shortShift128Left( aSig0, aSig1, expDiff, &aSig0, &aSig1 ); - } - mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); - sub128( aSig0, aSig1, term1, term2, &aSig0, &aSig1 ); - } - else { - shift128Right( aSig0, aSig1, 12, &aSig0, &aSig1 ); - shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 ); - } - do { - alternateASig0 = aSig0; - alternateASig1 = aSig1; - ++q; - sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 ); - } while ( 0 <= (sbits64) aSig0 ); - add128( - aSig0, aSig1, alternateASig0, alternateASig1, (bits64 *)&sigMean0, &sigMean1 ); - if ( ( sigMean0 < 0 ) - || ( ( ( sigMean0 | sigMean1 ) == 0 ) && ( q & 1 ) ) ) { - aSig0 = alternateASig0; - aSig1 = alternateASig1; - } - zSign = ( (sbits64) aSig0 < 0 ); - if ( zSign ) sub128( 0, 0, aSig0, aSig1, &aSig0, &aSig1 ); - return - normalizeRoundAndPackFloat128( aSign ^ zSign, bExp - 4, aSig0, aSig1 ); + if ( aExp == 0x7FFF ) { + if ( ( aSig0 | aSig1 ) + || ( ( bExp == 0x7FFF ) && ( bSig0 | bSig1 ) ) ) { + return propagateFloat128NaN( a, b ); + } + goto invalid; + } + if ( bExp == 0x7FFF ) { + if ( bSig0 | bSig1 ) return propagateFloat128NaN( a, b ); + return a; + } + if ( bExp == 0 ) { + if ( ( bSig0 | bSig1 ) == 0 ) { + invalid: + float_raise( float_flag_invalid ); + z.low = float128_default_nan_low; + z.high = float128_default_nan_high; + return z; + } + normalizeFloat128Subnormal( bSig0, bSig1, &bExp, &bSig0, &bSig1 ); + } + if ( aExp == 0 ) { + if ( ( aSig0 | aSig1 ) == 0 ) return a; + normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); + } + expDiff = aExp - bExp; + if ( expDiff < -1 ) return a; + shortShift128Left( + aSig0 | LIT64( 0x0001000000000000 ), + aSig1, + 15 - ( expDiff < 0 ), + &aSig0, + &aSig1 + ); + shortShift128Left( + bSig0 | LIT64( 0x0001000000000000 ), bSig1, 15, &bSig0, &bSig1 ); + q = le128( bSig0, bSig1, aSig0, aSig1 ); + if ( q ) sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 ); + expDiff -= 64; + while ( 0 < expDiff ) { + q = estimateDiv128To64( aSig0, aSig1, bSig0 ); + q = ( 4 < q ) ? q - 4 : 0; + mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); + shortShift192Left( term0, term1, term2, 61, &term1, &term2, &allZero ); + shortShift128Left( aSig0, aSig1, 61, &aSig0, &allZero ); + sub128( aSig0, 0, term1, term2, &aSig0, &aSig1 ); + expDiff -= 61; + } + if ( -64 < expDiff ) { + q = estimateDiv128To64( aSig0, aSig1, bSig0 ); + q = ( 4 < q ) ? q - 4 : 0; + q >>= - expDiff; + shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 ); + expDiff += 52; + if ( expDiff < 0 ) { + shift128Right( aSig0, aSig1, - expDiff, &aSig0, &aSig1 ); + } + else { + shortShift128Left( aSig0, aSig1, expDiff, &aSig0, &aSig1 ); + } + mul128By64To192( bSig0, bSig1, q, &term0, &term1, &term2 ); + sub128( aSig0, aSig1, term1, term2, &aSig0, &aSig1 ); + } + else { + shift128Right( aSig0, aSig1, 12, &aSig0, &aSig1 ); + shift128Right( bSig0, bSig1, 12, &bSig0, &bSig1 ); + } + do { + alternateASig0 = aSig0; + alternateASig1 = aSig1; + ++q; + sub128( aSig0, aSig1, bSig0, bSig1, &aSig0, &aSig1 ); + } while ( 0 <= (sbits64) aSig0 ); + add128( + aSig0, aSig1, alternateASig0, alternateASig1, (bits64 *)&sigMean0, &sigMean1 ); + if ( ( sigMean0 < 0 ) + || ( ( ( sigMean0 | sigMean1 ) == 0 ) && ( q & 1 ) ) ) { + aSig0 = alternateASig0; + aSig1 = alternateASig1; + } + zSign = ( (sbits64) aSig0 < 0 ); + if ( zSign ) sub128( 0, 0, aSig0, aSig1, &aSig0, &aSig1 ); + return + normalizeRoundAndPackFloat128( aSign ^ zSign, bExp - 4, aSig0, aSig1 ); } @@ -4704,64 +4704,64 @@ float128 float128_rem( float128 a, float128 b ) float128 float128_sqrt( float128 a ) { - flag aSign; - int32 aExp, zExp; - bits64 aSig0, aSig1, zSig0, zSig1, zSig2, doubleZSig0; - bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; - float128 z; - - aSig1 = extractFloat128Frac1( a ); - aSig0 = extractFloat128Frac0( a ); - aExp = extractFloat128Exp( a ); - aSign = extractFloat128Sign( a ); - if ( aExp == 0x7FFF ) { - if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, a ); - if ( ! aSign ) return a; - goto invalid; - } - if ( aSign ) { - if ( ( aExp | aSig0 | aSig1 ) == 0 ) return a; - invalid: - float_raise( float_flag_invalid ); - z.low = float128_default_nan_low; - z.high = float128_default_nan_high; - return z; - } - if ( aExp == 0 ) { - if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( 0, 0, 0, 0 ); - normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); - } - zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFE; - aSig0 |= LIT64( 0x0001000000000000 ); - zSig0 = estimateSqrt32( aExp, aSig0>>17 ); - shortShift128Left( aSig0, aSig1, 13 - ( aExp & 1 ), &aSig0, &aSig1 ); - zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); - doubleZSig0 = zSig0<<1; - mul64To128( zSig0, zSig0, &term0, &term1 ); - sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); - while ( (sbits64) rem0 < 0 ) { - --zSig0; - doubleZSig0 -= 2; - add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); - } - zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 ); - if ( ( zSig1 & 0x1FFF ) <= 5 ) { - if ( zSig1 == 0 ) zSig1 = 1; - mul64To128( doubleZSig0, zSig1, &term1, &term2 ); - sub128( rem1, 0, term1, term2, &rem1, &rem2 ); - mul64To128( zSig1, zSig1, &term2, &term3 ); - sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); - while ( (sbits64) rem1 < 0 ) { - --zSig1; - shortShift128Left( 0, zSig1, 1, &term2, &term3 ); - term3 |= 1; - term2 |= doubleZSig0; - add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 ); - } - zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); - } - shift128ExtraRightJamming( zSig0, zSig1, 0, 14, &zSig0, &zSig1, &zSig2 ); - return roundAndPackFloat128( 0, zExp, zSig0, zSig1, zSig2 ); + flag aSign; + int32 aExp, zExp; + bits64 aSig0, aSig1, zSig0, zSig1, zSig2, doubleZSig0; + bits64 rem0, rem1, rem2, rem3, term0, term1, term2, term3; + float128 z; + + aSig1 = extractFloat128Frac1( a ); + aSig0 = extractFloat128Frac0( a ); + aExp = extractFloat128Exp( a ); + aSign = extractFloat128Sign( a ); + if ( aExp == 0x7FFF ) { + if ( aSig0 | aSig1 ) return propagateFloat128NaN( a, a ); + if ( ! aSign ) return a; + goto invalid; + } + if ( aSign ) { + if ( ( aExp | aSig0 | aSig1 ) == 0 ) return a; + invalid: + float_raise( float_flag_invalid ); + z.low = float128_default_nan_low; + z.high = float128_default_nan_high; + return z; + } + if ( aExp == 0 ) { + if ( ( aSig0 | aSig1 ) == 0 ) return packFloat128( 0, 0, 0, 0 ); + normalizeFloat128Subnormal( aSig0, aSig1, &aExp, &aSig0, &aSig1 ); + } + zExp = ( ( aExp - 0x3FFF )>>1 ) + 0x3FFE; + aSig0 |= LIT64( 0x0001000000000000 ); + zSig0 = estimateSqrt32( aExp, aSig0>>17 ); + shortShift128Left( aSig0, aSig1, 13 - ( aExp & 1 ), &aSig0, &aSig1 ); + zSig0 = estimateDiv128To64( aSig0, aSig1, zSig0<<32 ) + ( zSig0<<30 ); + doubleZSig0 = zSig0<<1; + mul64To128( zSig0, zSig0, &term0, &term1 ); + sub128( aSig0, aSig1, term0, term1, &rem0, &rem1 ); + while ( (sbits64) rem0 < 0 ) { + --zSig0; + doubleZSig0 -= 2; + add128( rem0, rem1, zSig0>>63, doubleZSig0 | 1, &rem0, &rem1 ); + } + zSig1 = estimateDiv128To64( rem1, 0, doubleZSig0 ); + if ( ( zSig1 & 0x1FFF ) <= 5 ) { + if ( zSig1 == 0 ) zSig1 = 1; + mul64To128( doubleZSig0, zSig1, &term1, &term2 ); + sub128( rem1, 0, term1, term2, &rem1, &rem2 ); + mul64To128( zSig1, zSig1, &term2, &term3 ); + sub192( rem1, rem2, 0, 0, term2, term3, &rem1, &rem2, &rem3 ); + while ( (sbits64) rem1 < 0 ) { + --zSig1; + shortShift128Left( 0, zSig1, 1, &term2, &term3 ); + term3 |= 1; + term2 |= doubleZSig0; + add192( rem1, rem2, rem3, 0, term2, term3, &rem1, &rem2, &rem3 ); + } + zSig1 |= ( ( rem1 | rem2 | rem3 ) != 0 ); + } + shift128ExtraRightJamming( zSig0, zSig1, 0, 14, &zSig0, &zSig1, &zSig2 ); + return roundAndPackFloat128( 0, zExp, zSig0, zSig1, zSig2 ); } @@ -4774,23 +4774,23 @@ float128 float128_sqrt( float128 a ) flag float128_eq( float128 a, float128 b ) { - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - if ( float128_is_signaling_nan( a ) - || float128_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - return - ( a.low == b.low ) - && ( ( a.high == b.high ) - || ( ( a.low == 0 ) - && ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) - ); + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + if ( float128_is_signaling_nan( a ) + || float128_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + return + ( a.low == b.low ) + && ( ( a.high == b.high ) + || ( ( a.low == 0 ) + && ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) + ); } @@ -4803,27 +4803,27 @@ flag float128_eq( float128 a, float128 b ) flag float128_le( float128 a, float128 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign != bSign ) { - return - aSign - || ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - == 0 ); - } - return - aSign ? le128( b.high, b.low, a.high, a.low ) - : le128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign != bSign ) { + return + aSign + || ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + == 0 ); + } + return + aSign ? le128( b.high, b.low, a.high, a.low ) + : le128( a.high, a.low, b.high, b.low ); } @@ -4835,27 +4835,27 @@ flag float128_le( float128 a, float128 b ) flag float128_lt( float128 a, float128 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign != bSign ) { - return - aSign - && ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - != 0 ); - } - return - aSign ? lt128( b.high, b.low, a.high, a.low ) - : lt128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign != bSign ) { + return + aSign + && ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + != 0 ); + } + return + aSign ? lt128( b.high, b.low, a.high, a.low ) + : lt128( a.high, a.low, b.high, b.low ); } @@ -4869,20 +4869,20 @@ flag float128_lt( float128 a, float128 b ) flag float128_eq_signaling( float128 a, float128 b ) { - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - float_raise( float_flag_invalid ); - return 0; - } - return - ( a.low == b.low ) - && ( ( a.high == b.high ) - || ( ( a.low == 0 ) - && ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) - ); + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + float_raise( float_flag_invalid ); + return 0; + } + return + ( a.low == b.low ) + && ( ( a.high == b.high ) + || ( ( a.low == 0 ) + && ( (bits64) ( ( a.high | b.high )<<1 ) == 0 ) ) + ); } @@ -4895,30 +4895,30 @@ flag float128_eq_signaling( float128 a, float128 b ) flag float128_le_quiet( float128 a, float128 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - if ( float128_is_signaling_nan( a ) - || float128_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign != bSign ) { - return - aSign - || ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - == 0 ); - } - return - aSign ? le128( b.high, b.low, a.high, a.low ) - : le128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + if ( float128_is_signaling_nan( a ) + || float128_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign != bSign ) { + return + aSign + || ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + == 0 ); + } + return + aSign ? le128( b.high, b.low, a.high, a.low ) + : le128( a.high, a.low, b.high, b.low ); } @@ -4931,32 +4931,31 @@ flag float128_le_quiet( float128 a, float128 b ) flag float128_lt_quiet( float128 a, float128 b ) { - flag aSign, bSign; - - if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) - && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) - || ( ( extractFloat128Exp( b ) == 0x7FFF ) - && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) - ) { - if ( float128_is_signaling_nan( a ) - || float128_is_signaling_nan( b ) ) { - float_raise( float_flag_invalid ); - } - return 0; - } - aSign = extractFloat128Sign( a ); - bSign = extractFloat128Sign( b ); - if ( aSign != bSign ) { - return - aSign - && ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) - != 0 ); - } - return - aSign ? lt128( b.high, b.low, a.high, a.low ) - : lt128( a.high, a.low, b.high, b.low ); + flag aSign, bSign; + + if ( ( ( extractFloat128Exp( a ) == 0x7FFF ) + && ( extractFloat128Frac0( a ) | extractFloat128Frac1( a ) ) ) + || ( ( extractFloat128Exp( b ) == 0x7FFF ) + && ( extractFloat128Frac0( b ) | extractFloat128Frac1( b ) ) ) + ) { + if ( float128_is_signaling_nan( a ) + || float128_is_signaling_nan( b ) ) { + float_raise( float_flag_invalid ); + } + return 0; + } + aSign = extractFloat128Sign( a ); + bSign = extractFloat128Sign( b ); + if ( aSign != bSign ) { + return + aSign + && ( ( ( (bits64) ( ( a.high | b.high )<<1 ) ) | a.low | b.low ) + != 0 ); + } + return + aSign ? lt128( b.high, b.low, a.high, a.low ) + : lt128( a.high, a.low, b.high, b.low ); } #endif - diff --git a/src/lib/softfloat/softfloat.h b/src/lib/softfloat/softfloat.h index 9e664c85bc0..260c4810332 100644 --- a/src/lib/softfloat/softfloat.h +++ b/src/lib/softfloat/softfloat.h @@ -47,13 +47,13 @@ typedef bits32 float32; typedef bits64 float64; #ifdef FLOATX80 typedef struct { - bits16 high; - bits64 low; + bits16 high; + bits64 low; } floatx80; #endif #ifdef FLOAT128 typedef struct { - bits64 high, low; + bits64 high, low; } float128; #endif @@ -69,8 +69,8 @@ typedef struct { *----------------------------------------------------------------------------*/ extern int8 float_detect_tininess; enum { - float_tininess_after_rounding = 0, - float_tininess_before_rounding = 1 + float_tininess_after_rounding = 0, + float_tininess_before_rounding = 1 }; /*---------------------------------------------------------------------------- @@ -78,10 +78,10 @@ enum { *----------------------------------------------------------------------------*/ extern int8 float_rounding_mode; enum { - float_round_nearest_even = 0, - float_round_to_zero = 1, - float_round_down = 2, - float_round_up = 3 + float_round_nearest_even = 0, + float_round_to_zero = 1, + float_round_down = 2, + float_round_up = 3 }; /*---------------------------------------------------------------------------- @@ -89,11 +89,11 @@ enum { *----------------------------------------------------------------------------*/ extern int8 float_exception_flags; enum { - float_flag_inexact = 1, - float_flag_underflow = 2, - float_flag_overflow = 4, - float_flag_divbyzero = 8, - float_flag_invalid = 16 + float_flag_inexact = 1, + float_flag_underflow = 2, + float_flag_overflow = 4, + float_flag_divbyzero = 8, + float_flag_invalid = 16 }; /*---------------------------------------------------------------------------- @@ -210,11 +210,11 @@ float128 floatx80_to_float128( floatx80 ); INLINE floatx80 packFloatx80( flag zSign, int32 zExp, bits64 zSig ) { - floatx80 z; + floatx80 z; - z.low = zSig; - z.high = ( ( (bits16) zSign )<<15 ) + zExp; - return z; + z.low = zSig; + z.high = ( ( (bits16) zSign )<<15 ) + zExp; + return z; } @@ -303,13 +303,13 @@ flag float128_is_signaling_nan( float128 ); *----------------------------------------------------------------------------*/ INLINE float128 - packFloat128( flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 ) + packFloat128( flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 ) { - float128 z; + float128 z; - z.low = zSig1; - z.high = ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<48 ) + zSig0; - return z; + z.low = zSig1; + z.high = ( ( (bits64) zSign )<<63 ) + ( ( (bits64) zExp )<<48 ) + zSig0; + return z; } @@ -335,92 +335,92 @@ INLINE float128 *----------------------------------------------------------------------------*/ INLINE float128 - roundAndPackFloat128( - flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1, bits64 zSig2 ) + roundAndPackFloat128( + flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1, bits64 zSig2 ) { - int8 roundingMode; - flag roundNearestEven, increment, isTiny; - - roundingMode = float_rounding_mode; - roundNearestEven = ( roundingMode == float_round_nearest_even ); - increment = ( (sbits64) zSig2 < 0 ); - if ( ! roundNearestEven ) { - if ( roundingMode == float_round_to_zero ) { - increment = 0; - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig2; - } - else { - increment = ( roundingMode == float_round_up ) && zSig2; - } - } - } - if ( 0x7FFD <= (bits32) zExp ) { - if ( ( 0x7FFD < zExp ) - || ( ( zExp == 0x7FFD ) - && eq128( - LIT64( 0x0001FFFFFFFFFFFF ), - LIT64( 0xFFFFFFFFFFFFFFFF ), - zSig0, - zSig1 - ) - && increment - ) - ) { - float_raise( float_flag_overflow | float_flag_inexact ); - if ( ( roundingMode == float_round_to_zero ) - || ( zSign && ( roundingMode == float_round_up ) ) - || ( ! zSign && ( roundingMode == float_round_down ) ) - ) { - return - packFloat128( - zSign, - 0x7FFE, - LIT64( 0x0000FFFFFFFFFFFF ), - LIT64( 0xFFFFFFFFFFFFFFFF ) - ); - } - return packFloat128( zSign, 0x7FFF, 0, 0 ); - } - if ( zExp < 0 ) { - isTiny = - ( float_detect_tininess == float_tininess_before_rounding ) - || ( zExp < -1 ) - || ! increment - || lt128( - zSig0, - zSig1, - LIT64( 0x0001FFFFFFFFFFFF ), - LIT64( 0xFFFFFFFFFFFFFFFF ) - ); - shift128ExtraRightJamming( - zSig0, zSig1, zSig2, - zExp, &zSig0, &zSig1, &zSig2 ); - zExp = 0; - if ( isTiny && zSig2 ) float_raise( float_flag_underflow ); - if ( roundNearestEven ) { - increment = ( (sbits64) zSig2 < 0 ); - } - else { - if ( zSign ) { - increment = ( roundingMode == float_round_down ) && zSig2; - } - else { - increment = ( roundingMode == float_round_up ) && zSig2; - } - } - } - } - if ( zSig2 ) float_exception_flags |= float_flag_inexact; - if ( increment ) { - add128( zSig0, zSig1, 0, 1, &zSig0, &zSig1 ); - zSig1 &= ~ ( ( zSig2 + zSig2 == 0 ) & roundNearestEven ); - } - else { - if ( ( zSig0 | zSig1 ) == 0 ) zExp = 0; - } - return packFloat128( zSign, zExp, zSig0, zSig1 ); + int8 roundingMode; + flag roundNearestEven, increment, isTiny; + + roundingMode = float_rounding_mode; + roundNearestEven = ( roundingMode == float_round_nearest_even ); + increment = ( (sbits64) zSig2 < 0 ); + if ( ! roundNearestEven ) { + if ( roundingMode == float_round_to_zero ) { + increment = 0; + } + else { + if ( zSign ) { + increment = ( roundingMode == float_round_down ) && zSig2; + } + else { + increment = ( roundingMode == float_round_up ) && zSig2; + } + } + } + if ( 0x7FFD <= (bits32) zExp ) { + if ( ( 0x7FFD < zExp ) + || ( ( zExp == 0x7FFD ) + && eq128( + LIT64( 0x0001FFFFFFFFFFFF ), + LIT64( 0xFFFFFFFFFFFFFFFF ), + zSig0, + zSig1 + ) + && increment + ) + ) { + float_raise( float_flag_overflow | float_flag_inexact ); + if ( ( roundingMode == float_round_to_zero ) + || ( zSign && ( roundingMode == float_round_up ) ) + || ( ! zSign && ( roundingMode == float_round_down ) ) + ) { + return + packFloat128( + zSign, + 0x7FFE, + LIT64( 0x0000FFFFFFFFFFFF ), + LIT64( 0xFFFFFFFFFFFFFFFF ) + ); + } + return packFloat128( zSign, 0x7FFF, 0, 0 ); + } + if ( zExp < 0 ) { + isTiny = + ( float_detect_tininess == float_tininess_before_rounding ) + || ( zExp < -1 ) + || ! increment + || lt128( + zSig0, + zSig1, + LIT64( 0x0001FFFFFFFFFFFF ), + LIT64( 0xFFFFFFFFFFFFFFFF ) + ); + shift128ExtraRightJamming( + zSig0, zSig1, zSig2, - zExp, &zSig0, &zSig1, &zSig2 ); + zExp = 0; + if ( isTiny && zSig2 ) float_raise( float_flag_underflow ); + if ( roundNearestEven ) { + increment = ( (sbits64) zSig2 < 0 ); + } + else { + if ( zSign ) { + increment = ( roundingMode == float_round_down ) && zSig2; + } + else { + increment = ( roundingMode == float_round_up ) && zSig2; + } + } + } + } + if ( zSig2 ) float_exception_flags |= float_flag_inexact; + if ( increment ) { + add128( zSig0, zSig1, 0, 1, &zSig0, &zSig1 ); + zSig1 &= ~ ( ( zSig2 + zSig2 == 0 ) & roundNearestEven ); + } + else { + if ( ( zSig0 | zSig1 ) == 0 ) zExp = 0; + } + return packFloat128( zSign, zExp, zSig0, zSig1 ); } @@ -435,30 +435,29 @@ INLINE float128 *----------------------------------------------------------------------------*/ INLINE float128 - normalizeRoundAndPackFloat128( - flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 ) + normalizeRoundAndPackFloat128( + flag zSign, int32 zExp, bits64 zSig0, bits64 zSig1 ) { - int8 shiftCount; - bits64 zSig2; - - if ( zSig0 == 0 ) { - zSig0 = zSig1; - zSig1 = 0; - zExp -= 64; - } - shiftCount = countLeadingZeros64( zSig0 ) - 15; - if ( 0 <= shiftCount ) { - zSig2 = 0; - shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); - } - else { - shift128ExtraRightJamming( - zSig0, zSig1, 0, - shiftCount, &zSig0, &zSig1, &zSig2 ); - } - zExp -= shiftCount; - return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); + int8 shiftCount; + bits64 zSig2; + + if ( zSig0 == 0 ) { + zSig0 = zSig1; + zSig1 = 0; + zExp -= 64; + } + shiftCount = countLeadingZeros64( zSig0 ) - 15; + if ( 0 <= shiftCount ) { + zSig2 = 0; + shortShift128Left( zSig0, zSig1, shiftCount, &zSig0, &zSig1 ); + } + else { + shift128ExtraRightJamming( + zSig0, zSig1, 0, - shiftCount, &zSig0, &zSig1, &zSig2 ); + } + zExp -= shiftCount; + return roundAndPackFloat128( zSign, zExp, zSig0, zSig1, zSig2 ); } #endif - |
