mirrors/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

softfloat: Convert float32_exp2 to FloatParts

Browse Source 
Keep the intermediate results in FloatParts instead of
converting back and forth between float64.  Use muladd
instead of separate mul+add.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This commit is contained in:
Richard Henderson 2020-11-21 19:20:23 -08:00
parent 1b96b006d2
commit 572c4d862f
1 changed files with 23 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

53
fpu/softfloat.c
View File

@ -5210,47 +5210,40 @@ static const float64 float32_exp2_coefficients[15] =
float32 float32_exp2(float32 a, float_status *status) float32 float32_exp2(float32 a, float_status *status)
{ {
bool aSign; FloatParts64 xp, xnp, tp, rp;
int aExp;
uint32_t aSig;
float64 r, x, xn;
int i; int i;
a = float32_squash_input_denormal(a, status);
aSig = extractFloat32Frac( a ); float32_unpack_canonical(&xp, a, status);
aExp = extractFloat32Exp( a ); if (unlikely(xp.cls != float_class_normal)) {
aSign = extractFloat32Sign( a ); switch (xp.cls) {
case float_class_snan:
if ( aExp == 0xFF) { case float_class_qnan:
if (aSig) { parts_return_nan(&xp, status);
return propagateFloat32NaN(a, float32_zero, status); return float32_round_pack_canonical(&xp, status);
case float_class_inf:
return xp.sign ? float32_zero : a;
case float_class_zero:
return float32_one;
default:
break;
} }
return (aSign) ? float32_zero : a; g_assert_not_reached();
}
if (aExp == 0) {
if (aSig == 0) return float32_one;
} }
float_raise(float_flag_inexact, status); float_raise(float_flag_inexact, status);
/* ******************************* */ float64_unpack_canonical(&xnp, float64_ln2, status);
/* using float64 for approximation */ xp = *parts_mul(&xp, &tp, status);
/* ******************************* */ xnp = xp;
x = float32_to_float64(a, status);
x = float64_mul(x, float64_ln2, status);
xn = x; float64_unpack_canonical(&rp, float64_one, status);
r = float64_one;
for (i = 0 ; i < 15 ; i++) { for (i = 0 ; i < 15 ; i++) {
float64 f; float64_unpack_canonical(&tp, float32_exp2_coefficients[i], status);
rp = *parts_muladd(&tp, &xp, &rp, 0, status);
f = float64_mul(xn, float32_exp2_coefficients[i], status); xnp = *parts_mul(&xnp, &xp, status);
r = float64_add(r, f, status);
xn = float64_mul(xn, x, status);
} }
return float64_to_float32(r, status); return float32_round_pack_canonical(&rp, status);
} }
/*---------------------------------------------------------------------------- /*----------------------------------------------------------------------------