diff options
author | Richard Henderson <richard.henderson@linaro.org> | 2020-10-24 06:04:19 -0700 |
---|---|---|
committer | Richard Henderson <richard.henderson@linaro.org> | 2021-05-16 07:13:51 -0500 |
commit | dedd123c56214f897d7045f2133b0261502690c6 (patch) | |
tree | c2dfb989e290a8bdb01d4599dfa7848c75cbfae9 /fpu | |
parent | aca845275a62e79daee8ed5bf95ccb8ace4aeac9 (diff) |
softfloat: Move muladd_floats to softfloat-parts.c.inc
Rename to parts$N_muladd.
Implement float128_muladd with FloatParts128.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Diffstat (limited to 'fpu')
-rw-r--r-- | fpu/softfloat-parts.c.inc | 126 | ||||
-rw-r--r-- | fpu/softfloat.c | 406 |
2 files changed, 323 insertions, 209 deletions
diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index 9a67ab2bea..a203811299 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -413,3 +413,129 @@ static FloatPartsN *partsN(mul)(FloatPartsN *a, FloatPartsN *b, a->sign = sign; return a; } + +/* + * Returns the result of multiplying the floating-point values `a' and + * `b' then adding 'c', with no intermediate rounding step after the + * multiplication. The operation is performed according to the + * IEC/IEEE Standard for Binary Floating-Point Arithmetic 754-2008. + * The flags argument allows the caller to select negation of the + * addend, the intermediate product, or the final result. (The + * difference between this and having the caller do a separate + * negation is that negating externally will flip the sign bit on NaNs.) + * + * Requires A and C extracted into a double-sized structure to provide the + * extra space for the widening multiply. + */ +static FloatPartsN *partsN(muladd)(FloatPartsN *a, FloatPartsN *b, + FloatPartsN *c, int flags, float_status *s) +{ + int ab_mask, abc_mask; + FloatPartsW p_widen, c_widen; + + ab_mask = float_cmask(a->cls) | float_cmask(b->cls); + abc_mask = float_cmask(c->cls) | ab_mask; + + /* + * It is implementation-defined whether the cases of (0,inf,qnan) + * and (inf,0,qnan) raise InvalidOperation or not (and what QNaN + * they return if they do), so we have to hand this information + * off to the target-specific pick-a-NaN routine. + */ + if (unlikely(abc_mask & float_cmask_anynan)) { + return parts_pick_nan_muladd(a, b, c, s, ab_mask, abc_mask); + } + + if (flags & float_muladd_negate_c) { + c->sign ^= 1; + } + + /* Compute the sign of the product into A. */ + a->sign ^= b->sign; + if (flags & float_muladd_negate_product) { + a->sign ^= 1; + } + + if (unlikely(ab_mask != float_cmask_normal)) { + if (unlikely(ab_mask == float_cmask_infzero)) { + goto d_nan; + } + + if (ab_mask & float_cmask_inf) { + if (c->cls == float_class_inf && a->sign != c->sign) { + goto d_nan; + } + goto return_inf; + } + + g_assert(ab_mask & float_cmask_zero); + if (c->cls == float_class_normal) { + *a = *c; + goto return_normal; + } + if (c->cls == float_class_zero) { + if (a->sign != c->sign) { + goto return_sub_zero; + } + goto return_zero; + } + g_assert(c->cls == float_class_inf); + } + + if (unlikely(c->cls == float_class_inf)) { + a->sign = c->sign; + goto return_inf; + } + + /* Perform the multiplication step. */ + p_widen.sign = a->sign; + p_widen.exp = a->exp + b->exp + 1; + frac_mulw(&p_widen, a, b); + if (!(p_widen.frac_hi & DECOMPOSED_IMPLICIT_BIT)) { + frac_add(&p_widen, &p_widen, &p_widen); + p_widen.exp -= 1; + } + + /* Perform the addition step. */ + if (c->cls != float_class_zero) { + /* Zero-extend C to less significant bits. */ + frac_widen(&c_widen, c); + c_widen.exp = c->exp; + + if (a->sign == c->sign) { + parts_add_normal(&p_widen, &c_widen); + } else if (!parts_sub_normal(&p_widen, &c_widen)) { + goto return_sub_zero; + } + } + + /* Narrow with sticky bit, for proper rounding later. */ + frac_truncjam(a, &p_widen); + a->sign = p_widen.sign; + a->exp = p_widen.exp; + + return_normal: + if (flags & float_muladd_halve_result) { + a->exp -= 1; + } + finish_sign: + if (flags & float_muladd_negate_result) { + a->sign ^= 1; + } + return a; + + return_sub_zero: + a->sign = s->float_rounding_mode == float_round_down; + return_zero: + a->cls = float_class_zero; + goto finish_sign; + + return_inf: + a->cls = float_class_inf; + goto finish_sign; + + d_nan: + float_raise(float_flag_invalid, s); + parts_default_nan(a, s); + return a; +} diff --git a/fpu/softfloat.c b/fpu/softfloat.c index ac7959557c..571309e74f 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -716,6 +716,10 @@ static float128 float128_pack_raw(const FloatParts128 *p) #define PARTS_GENERIC_64_128(NAME, P) \ QEMU_GENERIC(P, (FloatParts128 *, parts128_##NAME), parts64_##NAME) +#define PARTS_GENERIC_64_128_256(NAME, P) \ + QEMU_GENERIC(P, (FloatParts256 *, parts256_##NAME), \ + (FloatParts128 *, parts128_##NAME), parts64_##NAME) + #define parts_default_nan(P, S) PARTS_GENERIC_64_128(default_nan, P)(P, S) #define parts_silence_nan(P, S) PARTS_GENERIC_64_128(silence_nan, P)(P, S) @@ -761,15 +765,17 @@ static void parts128_uncanon(FloatParts128 *p, float_status *status, static void parts64_add_normal(FloatParts64 *a, FloatParts64 *b); static void parts128_add_normal(FloatParts128 *a, FloatParts128 *b); +static void parts256_add_normal(FloatParts256 *a, FloatParts256 *b); #define parts_add_normal(A, B) \ - PARTS_GENERIC_64_128(add_normal, A)(A, B) + PARTS_GENERIC_64_128_256(add_normal, A)(A, B) static bool parts64_sub_normal(FloatParts64 *a, FloatParts64 *b); static bool parts128_sub_normal(FloatParts128 *a, FloatParts128 *b); +static bool parts256_sub_normal(FloatParts256 *a, FloatParts256 *b); #define parts_sub_normal(A, B) \ - PARTS_GENERIC_64_128(sub_normal, A)(A, B) + PARTS_GENERIC_64_128_256(sub_normal, A)(A, B) static FloatParts64 *parts64_addsub(FloatParts64 *a, FloatParts64 *b, float_status *s, bool subtract); @@ -787,6 +793,16 @@ static FloatParts128 *parts128_mul(FloatParts128 *a, FloatParts128 *b, #define parts_mul(A, B, S) \ PARTS_GENERIC_64_128(mul, A)(A, B, S) +static FloatParts64 *parts64_muladd(FloatParts64 *a, FloatParts64 *b, + FloatParts64 *c, int flags, + float_status *s); +static FloatParts128 *parts128_muladd(FloatParts128 *a, FloatParts128 *b, + FloatParts128 *c, int flags, + float_status *s); + +#define parts_muladd(A, B, C, Z, S) \ + PARTS_GENERIC_64_128(muladd, A)(A, B, C, Z, S) + /* * Helper functions for softfloat-parts.c.inc, per-size operations. */ @@ -794,6 +810,10 @@ static FloatParts128 *parts128_mul(FloatParts128 *a, FloatParts128 *b, #define FRAC_GENERIC_64_128(NAME, P) \ QEMU_GENERIC(P, (FloatParts128 *, frac128_##NAME), frac64_##NAME) +#define FRAC_GENERIC_64_128_256(NAME, P) \ + QEMU_GENERIC(P, (FloatParts256 *, frac256_##NAME), \ + (FloatParts128 *, frac128_##NAME), frac64_##NAME) + static bool frac64_add(FloatParts64 *r, FloatParts64 *a, FloatParts64 *b) { return uadd64_overflow(a->frac, b->frac, &r->frac); @@ -807,7 +827,17 @@ static bool frac128_add(FloatParts128 *r, FloatParts128 *a, FloatParts128 *b) return c; } -#define frac_add(R, A, B) FRAC_GENERIC_64_128(add, R)(R, A, B) +static bool frac256_add(FloatParts256 *r, FloatParts256 *a, FloatParts256 *b) +{ + bool c = 0; + r->frac_lo = uadd64_carry(a->frac_lo, b->frac_lo, &c); + r->frac_lm = uadd64_carry(a->frac_lm, b->frac_lm, &c); + r->frac_hm = uadd64_carry(a->frac_hm, b->frac_hm, &c); + r->frac_hi = uadd64_carry(a->frac_hi, b->frac_hi, &c); + return c; +} + +#define frac_add(R, A, B) FRAC_GENERIC_64_128_256(add, R)(R, A, B) static bool frac64_addi(FloatParts64 *r, FloatParts64 *a, uint64_t c) { @@ -902,7 +932,16 @@ static void frac128_neg(FloatParts128 *a) a->frac_hi = usub64_borrow(0, a->frac_hi, &c); } -#define frac_neg(A) FRAC_GENERIC_64_128(neg, A)(A) +static void frac256_neg(FloatParts256 *a) +{ + bool c = 0; + a->frac_lo = usub64_borrow(0, a->frac_lo, &c); + a->frac_lm = usub64_borrow(0, a->frac_lm, &c); + a->frac_hm = usub64_borrow(0, a->frac_hm, &c); + a->frac_hi = usub64_borrow(0, a->frac_hi, &c); +} + +#define frac_neg(A) FRAC_GENERIC_64_128_256(neg, A)(A) static int frac64_normalize(FloatParts64 *a) { @@ -933,7 +972,55 @@ static int frac128_normalize(FloatParts128 *a) return 128; } -#define frac_normalize(A) FRAC_GENERIC_64_128(normalize, A)(A) +static int frac256_normalize(FloatParts256 *a) +{ + uint64_t a0 = a->frac_hi, a1 = a->frac_hm; + uint64_t a2 = a->frac_lm, a3 = a->frac_lo; + int ret, shl, shr; + + if (likely(a0)) { + shl = clz64(a0); + if (shl == 0) { + return 0; + } + ret = shl; + } else { + if (a1) { + ret = 64; + a0 = a1, a1 = a2, a2 = a3, a3 = 0; + } else if (a2) { + ret = 128; + a0 = a2, a1 = a3, a2 = 0, a3 = 0; + } else if (a3) { + ret = 192; + a0 = a3, a1 = 0, a2 = 0, a3 = 0; + } else { + ret = 256; + a0 = 0, a1 = 0, a2 = 0, a3 = 0; + goto done; + } + shl = clz64(a0); + if (shl == 0) { + goto done; + } + ret += shl; + } + + shr = -shl & 63; + a0 = (a0 << shl) | (a1 >> shr); + a1 = (a1 << shl) | (a2 >> shr); + a2 = (a2 << shl) | (a3 >> shr); + a3 = (a3 << shl); + + done: + a->frac_hi = a0; + a->frac_hm = a1; + a->frac_lm = a2; + a->frac_lo = a3; + return ret; +} + +#define frac_normalize(A) FRAC_GENERIC_64_128_256(normalize, A)(A) static void frac64_shl(FloatParts64 *a, int c) { @@ -969,7 +1056,51 @@ static void frac128_shrjam(FloatParts128 *a, int c) shift128RightJamming(a->frac_hi, a->frac_lo, c, &a->frac_hi, &a->frac_lo); } -#define frac_shrjam(A, C) FRAC_GENERIC_64_128(shrjam, A)(A, C) +static void frac256_shrjam(FloatParts256 *a, int c) +{ + uint64_t a0 = a->frac_hi, a1 = a->frac_hm; + uint64_t a2 = a->frac_lm, a3 = a->frac_lo; + uint64_t sticky = 0; + int invc; + + if (unlikely(c == 0)) { + return; + } else if (likely(c < 64)) { + /* nothing */ + } else if (likely(c < 256)) { + if (unlikely(c & 128)) { + sticky |= a2 | a3; + a3 = a1, a2 = a0, a1 = 0, a0 = 0; + } + if (unlikely(c & 64)) { + sticky |= a3; + a3 = a2, a2 = a1, a1 = a0, a0 = 0; + } + c &= 63; + if (c == 0) { + goto done; + } + } else { + sticky = a0 | a1 | a2 | a3; + a0 = a1 = a2 = a3 = 0; + goto done; + } + + invc = -c & 63; + sticky |= a3 << invc; + a3 = (a3 >> c) | (a2 << invc); + a2 = (a2 >> c) | (a1 << invc); + a1 = (a1 >> c) | (a0 << invc); + a0 = (a0 >> c); + + done: + a->frac_lo = a3 | (sticky != 0); + a->frac_lm = a2; + a->frac_hm = a1; + a->frac_hi = a0; +} + +#define frac_shrjam(A, C) FRAC_GENERIC_64_128_256(shrjam, A)(A, C) static bool frac64_sub(FloatParts64 *r, FloatParts64 *a, FloatParts64 *b) { @@ -984,7 +1115,17 @@ static bool frac128_sub(FloatParts128 *r, FloatParts128 *a, FloatParts128 *b) return c; } -#define frac_sub(R, A, B) FRAC_GENERIC_64_128(sub, R)(R, A, B) +static bool frac256_sub(FloatParts256 *r, FloatParts256 *a, FloatParts256 *b) +{ + bool c = 0; + r->frac_lo = usub64_borrow(a->frac_lo, b->frac_lo, &c); + r->frac_lm = usub64_borrow(a->frac_lm, b->frac_lm, &c); + r->frac_hm = usub64_borrow(a->frac_hm, b->frac_hm, &c); + r->frac_hi = usub64_borrow(a->frac_hi, b->frac_hi, &c); + return c; +} + +#define frac_sub(R, A, B) FRAC_GENERIC_64_128_256(sub, R)(R, A, B) static void frac64_truncjam(FloatParts64 *r, FloatParts128 *a) { @@ -999,6 +1140,22 @@ static void frac128_truncjam(FloatParts128 *r, FloatParts256 *a) #define frac_truncjam(R, A) FRAC_GENERIC_64_128(truncjam, R)(R, A) +static void frac64_widen(FloatParts128 *r, FloatParts64 *a) +{ + r->frac_hi = a->frac; + r->frac_lo = 0; +} + +static void frac128_widen(FloatParts256 *r, FloatParts128 *a) +{ + r->frac_hi = a->frac_hi; + r->frac_hm = a->frac_lo; + r->frac_lm = 0; + r->frac_lo = 0; +} + +#define frac_widen(A, B) FRAC_GENERIC_64_128(widen, B)(A, B) + #define partsN(NAME) glue(glue(glue(parts,N),_),NAME) #define FloatPartsN glue(FloatParts,N) #define FloatPartsW glue(FloatParts,W) @@ -1019,6 +1176,12 @@ static void frac128_truncjam(FloatParts128 *r, FloatParts256 *a) #undef N #undef W +#define N 256 + +#include "softfloat-parts-addsub.c.inc" + +#undef N +#undef W #undef partsN #undef FloatPartsN #undef FloatPartsW @@ -1387,230 +1550,48 @@ float128_mul(float128 a, float128 b, float_status *status) } /* - * Returns the result of multiplying the floating-point values `a' and - * `b' then adding 'c', with no intermediate rounding step after the - * multiplication. The operation is performed according to the - * IEC/IEEE Standard for Binary Floating-Point Arithmetic 754-2008. - * The flags argument allows the caller to select negation of the - * addend, the intermediate product, or the final result. (The - * difference between this and having the caller do a separate - * negation is that negating externally will flip the sign bit on - * NaNs.) + * Fused multiply-add */ -static FloatParts64 muladd_floats(FloatParts64 a, FloatParts64 b, FloatParts64 c, - int flags, float_status *s) -{ - bool inf_zero, p_sign; - bool sign_flip = flags & float_muladd_negate_result; - FloatClass p_class; - uint64_t hi, lo; - int p_exp; - int ab_mask, abc_mask; - - ab_mask = float_cmask(a.cls) | float_cmask(b.cls); - abc_mask = float_cmask(c.cls) | ab_mask; - inf_zero = ab_mask == float_cmask_infzero; - - /* It is implementation-defined whether the cases of (0,inf,qnan) - * and (inf,0,qnan) raise InvalidOperation or not (and what QNaN - * they return if they do), so we have to hand this information - * off to the target-specific pick-a-NaN routine. - */ - if (unlikely(abc_mask & float_cmask_anynan)) { - return *parts_pick_nan_muladd(&a, &b, &c, s, ab_mask, abc_mask); - } - - if (inf_zero) { - float_raise(float_flag_invalid, s); - parts_default_nan(&a, s); - return a; - } - - if (flags & float_muladd_negate_c) { - c.sign ^= 1; - } - - p_sign = a.sign ^ b.sign; - - if (flags & float_muladd_negate_product) { - p_sign ^= 1; - } - - if (ab_mask & float_cmask_inf) { - p_class = float_class_inf; - } else if (ab_mask & float_cmask_zero) { - p_class = float_class_zero; - } else { - p_class = float_class_normal; - } - - if (c.cls == float_class_inf) { - if (p_class == float_class_inf && p_sign != c.sign) { - float_raise(float_flag_invalid, s); - parts_default_nan(&c, s); - } else { - c.sign ^= sign_flip; - } - return c; - } - - if (p_class == float_class_inf) { - a.cls = float_class_inf; - a.sign = p_sign ^ sign_flip; - return a; - } - - if (p_class == float_class_zero) { - if (c.cls == float_class_zero) { - if (p_sign != c.sign) { - p_sign = s->float_rounding_mode == float_round_down; - } - c.sign = p_sign; - } else if (flags & float_muladd_halve_result) { - c.exp -= 1; - } - c.sign ^= sign_flip; - return c; - } - - /* a & b should be normals now... */ - assert(a.cls == float_class_normal && - b.cls == float_class_normal); - - p_exp = a.exp + b.exp; - - mul64To128(a.frac, b.frac, &hi, &lo); - - /* Renormalize to the msb. */ - if (hi & DECOMPOSED_IMPLICIT_BIT) { - p_exp += 1; - } else { - shortShift128Left(hi, lo, 1, &hi, &lo); - } - - /* + add/sub */ - if (c.cls != float_class_zero) { - int exp_diff = p_exp - c.exp; - if (p_sign == c.sign) { - /* Addition */ - if (exp_diff <= 0) { - shift64RightJamming(hi, -exp_diff, &hi); - p_exp = c.exp; - if (uadd64_overflow(hi, c.frac, &hi)) { - shift64RightJamming(hi, 1, &hi); - hi |= DECOMPOSED_IMPLICIT_BIT; - p_exp += 1; - } - } else { - uint64_t c_hi, c_lo, over; - shift128RightJamming(c.frac, 0, exp_diff, &c_hi, &c_lo); - add192(0, hi, lo, 0, c_hi, c_lo, &over, &hi, &lo); - if (over) { - shift64RightJamming(hi, 1, &hi); - hi |= DECOMPOSED_IMPLICIT_BIT; - p_exp += 1; - } - } - } else { - /* Subtraction */ - uint64_t c_hi = c.frac, c_lo = 0; - - if (exp_diff <= 0) { - shift128RightJamming(hi, lo, -exp_diff, &hi, &lo); - if (exp_diff == 0 - && - (hi > c_hi || (hi == c_hi && lo >= c_lo))) { - sub128(hi, lo, c_hi, c_lo, &hi, &lo); - } else { - sub128(c_hi, c_lo, hi, lo, &hi, &lo); - p_sign ^= 1; - p_exp = c.exp; - } - } else { - shift128RightJamming(c_hi, c_lo, - exp_diff, - &c_hi, &c_lo); - sub128(hi, lo, c_hi, c_lo, &hi, &lo); - } - - if (hi == 0 && lo == 0) { - a.cls = float_class_zero; - a.sign = s->float_rounding_mode == float_round_down; - a.sign ^= sign_flip; - return a; - } else { - int shift; - if (hi != 0) { - shift = clz64(hi); - } else { - shift = clz64(lo) + 64; - } - /* Normalizing to a binary point of 124 is the - correct adjust for the exponent. However since we're - shifting, we might as well put the binary point back - at 63 where we really want it. Therefore shift as - if we're leaving 1 bit at the top of the word, but - adjust the exponent as if we're leaving 3 bits. */ - shift128Left(hi, lo, shift, &hi, &lo); - p_exp -= shift; - } - } - } - hi |= (lo != 0); - - if (flags & float_muladd_halve_result) { - p_exp -= 1; - } - - /* finally prepare our result */ - a.cls = float_class_normal; - a.sign = p_sign ^ sign_flip; - a.exp = p_exp; - a.frac = hi; - - return a; -} - float16 QEMU_FLATTEN float16_muladd(float16 a, float16 b, float16 c, - int flags, float_status *status) + int flags, float_status *status) { - FloatParts64 pa, pb, pc, pr; + FloatParts64 pa, pb, pc, *pr; float16_unpack_canonical(&pa, a, status); float16_unpack_canonical(&pb, b, status); float16_unpack_canonical(&pc, c, status); - pr = muladd_floats(pa, pb, pc, flags, status); + pr = parts_muladd(&pa, &pb, &pc, flags, status); - return float16_round_pack_canonical(&pr, status); + return float16_round_pack_canonical(pr, status); } static float32 QEMU_SOFTFLOAT_ATTR soft_f32_muladd(float32 a, float32 b, float32 c, int flags, float_status *status) { - FloatParts64 pa, pb, pc, pr; + FloatParts64 pa, pb, pc, *pr; float32_unpack_canonical(&pa, a, status); float32_unpack_canonical(&pb, b, status); float32_unpack_canonical(&pc, c, status); - pr = muladd_floats(pa, pb, pc, flags, status); + pr = parts_muladd(&pa, &pb, &pc, flags, status); - return float32_round_pack_canonical(&pr, status); + return float32_round_pack_canonical(pr, status); } static float64 QEMU_SOFTFLOAT_ATTR soft_f64_muladd(float64 a, float64 b, float64 c, int flags, float_status *status) { - FloatParts64 pa, pb, pc, pr; + FloatParts64 pa, pb, pc, *pr; float64_unpack_canonical(&pa, a, status); float64_unpack_canonical(&pb, b, status); float64_unpack_canonical(&pc, c, status); - pr = muladd_floats(pa, pb, pc, flags, status); + pr = parts_muladd(&pa, &pb, &pc, flags, status); - return float64_round_pack_canonical(&pr, status); + return float64_round_pack_canonical(pr, status); } static bool force_soft_fma; @@ -1757,23 +1738,30 @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s) return soft_f64_muladd(ua.s, ub.s, uc.s, flags, s); } -/* - * Returns the result of multiplying the bfloat16 values `a' - * and `b' then adding 'c', with no intermediate rounding step after the - * multiplication. - */ - bfloat16 QEMU_FLATTEN bfloat16_muladd(bfloat16 a, bfloat16 b, bfloat16 c, int flags, float_status *status) { - FloatParts64 pa, pb, pc, pr; + FloatParts64 pa, pb, pc, *pr; bfloat16_unpack_canonical(&pa, a, status); bfloat16_unpack_canonical(&pb, b, status); bfloat16_unpack_canonical(&pc, c, status); - pr = muladd_floats(pa, pb, pc, flags, status); + pr = parts_muladd(&pa, &pb, &pc, flags, status); - return bfloat16_round_pack_canonical(&pr, status); + return bfloat16_round_pack_canonical(pr, status); +} + +float128 QEMU_FLATTEN float128_muladd(float128 a, float128 b, float128 c, + int flags, float_status *status) +{ + FloatParts128 pa, pb, pc, *pr; + + float128_unpack_canonical(&pa, a, status); + float128_unpack_canonical(&pb, b, status); + float128_unpack_canonical(&pc, c, status); + pr = parts_muladd(&pa, &pb, &pc, flags, status); + + return float128_round_pack_canonical(pr, status); } /* |