[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 29/29] softfloat: Use hard-float for {u}int64_to_float{32,64}
From: |
Richard Henderson |
Subject: |
[PULL 29/29] softfloat: Use hard-float for {u}int64_to_float{32,64} |
Date: |
Thu, 3 Jun 2021 14:41:31 -0700 |
For the normal case of no additional scaling, this reduces the
profile contribution of int64_to_float64 to the testcase in the
linked issue from 0.81% to 0.04%.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/134
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
fpu/softfloat.c | 28 ++++++++++++++++++++++++++++
1 file changed, 28 insertions(+)
diff --git a/fpu/softfloat.c b/fpu/softfloat.c
index 5026f518b0..1cb162882b 100644
--- a/fpu/softfloat.c
+++ b/fpu/softfloat.c
@@ -3559,6 +3559,13 @@ float32 int64_to_float32_scalbn(int64_t a, int scale,
float_status *status)
{
FloatParts64 p;
+ /* Without scaling, there are no overflow concerns. */
+ if (likely(scale == 0) && can_use_fpu(status)) {
+ union_float32 ur;
+ ur.h = a;
+ return ur.s;
+ }
+
parts64_sint_to_float(&p, a, scale, status);
return float32_round_pack_canonical(&p, status);
}
@@ -3592,6 +3599,13 @@ float64 int64_to_float64_scalbn(int64_t a, int scale,
float_status *status)
{
FloatParts64 p;
+ /* Without scaling, there are no overflow concerns. */
+ if (likely(scale == 0) && can_use_fpu(status)) {
+ union_float64 ur;
+ ur.h = a;
+ return ur.s;
+ }
+
parts_sint_to_float(&p, a, scale, status);
return float64_round_pack_canonical(&p, status);
}
@@ -3726,6 +3740,13 @@ float32 uint64_to_float32_scalbn(uint64_t a, int scale,
float_status *status)
{
FloatParts64 p;
+ /* Without scaling, there are no overflow concerns. */
+ if (likely(scale == 0) && can_use_fpu(status)) {
+ union_float32 ur;
+ ur.h = a;
+ return ur.s;
+ }
+
parts_uint_to_float(&p, a, scale, status);
return float32_round_pack_canonical(&p, status);
}
@@ -3759,6 +3780,13 @@ float64 uint64_to_float64_scalbn(uint64_t a, int scale,
float_status *status)
{
FloatParts64 p;
+ /* Without scaling, there are no overflow concerns. */
+ if (likely(scale == 0) && can_use_fpu(status)) {
+ union_float64 ur;
+ ur.h = a;
+ return ur.s;
+ }
+
parts_uint_to_float(&p, a, scale, status);
return float64_round_pack_canonical(&p, status);
}
--
2.25.1
- [PULL 19/29] softfloat: Convert floatx80_round_to_int to FloatParts, (continued)
- [PULL 19/29] softfloat: Convert floatx80_round_to_int to FloatParts, Richard Henderson, 2021/06/03
- [PULL 16/29] softfloat: Convert floatx80_div to FloatParts, Richard Henderson, 2021/06/03
- [PULL 22/29] softfloat: Convert floatx80 to integer to FloatParts, Richard Henderson, 2021/06/03
- [PULL 21/29] softfloat: Convert floatx80 float conversions to FloatParts, Richard Henderson, 2021/06/03
- [PULL 17/29] softfloat: Convert floatx80_sqrt to FloatParts, Richard Henderson, 2021/06/03
- [PULL 15/29] softfloat: Convert floatx80_mul to FloatParts, Richard Henderson, 2021/06/03
- [PULL 24/29] softfloat: Convert floatx80 compare to FloatParts, Richard Henderson, 2021/06/03
- [PULL 20/29] softfloat: Convert integer to floatx80 to FloatParts, Richard Henderson, 2021/06/03
- [PULL 23/29] softfloat: Convert floatx80_scalbn to FloatParts, Richard Henderson, 2021/06/03
- [PULL 28/29] tests/fp: Enable more tests, Richard Henderson, 2021/06/03
- [PULL 29/29] softfloat: Use hard-float for {u}int64_to_float{32,64},
Richard Henderson <=
- [PULL 26/29] softfloat: Move floatN_log2 to softfloat-parts.c.inc, Richard Henderson, 2021/06/03
- [PULL 25/29] softfloat: Convert float32_exp2 to FloatParts, Richard Henderson, 2021/06/03
- [PULL 27/29] softfloat: Convert modrem operations to FloatParts, Richard Henderson, 2021/06/03
- Re: [PULL 00/29] softfloat patch queue, Peter Maydell, 2021/06/04