qiurui
/
glibc
mirror of git://sourceware.org/git/glibc.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

AArch64: Remove -0.0 check from vector sin 

Remove the unnecessary extra checks for sin (-0.0) from vector sin/sinf,
improving performance.  Passes regress.

Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
This commit is contained in:
Wilco Dijkstra 2023-09-19 14:03:48 +01:00
parent fd134feba3
commit 6b695e5c62
2 changed files with 2 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
sysdeps/aarch64/fpu/sin_advsimd.c
View File

@ -56,7 +56,7 @@ float64x2_t VPCS_ATTR V_NAME_D1 (sin) (float64x2_t x)
{ {
const struct data *d = ptr_barrier (&data); const struct data *d = ptr_barrier (&data);
float64x2_t n, r, r2, r3, r4, y, t1, t2, t3; float64x2_t n, r, r2, r3, r4, y, t1, t2, t3;
uint64x2_t odd, cmp, eqz; uint64x2_t odd, cmp;
#if WANT_SIMD_EXCEPT #if WANT_SIMD_EXCEPT
/* Detect |x| <= TinyBound or |x| >= RangeVal. If fenv exceptions are to be /* Detect |x| <= TinyBound or |x| >= RangeVal. If fenv exceptions are to be
@ -70,7 +70,6 @@ float64x2_t VPCS_ATTR V_NAME_D1 (sin) (float64x2_t x)
cmp = vcageq_f64 (d->range_val, x); cmp = vcageq_f64 (d->range_val, x);
cmp = vceqzq_u64 (cmp); /* cmp = ~cmp. */ cmp = vceqzq_u64 (cmp); /* cmp = ~cmp. */
#endif #endif
eqz = vceqzq_f64 (x);
/* n = rint(|x|/pi). */ /* n = rint(|x|/pi). */
n = vfmaq_f64 (d->shift, d->inv_pi, r); n = vfmaq_f64 (d->shift, d->inv_pi, r);
@ -96,10 +95,6 @@ float64x2_t VPCS_ATTR V_NAME_D1 (sin) (float64x2_t x)
y = vfmaq_f64 (t3, y, r4); y = vfmaq_f64 (t3, y, r4);
y = vfmaq_f64 (r, y, r3); y = vfmaq_f64 (r, y, r3);
/* Sign of 0 is discarded by polynomial, so copy it back here. */
if (__glibc_unlikely (v_any_u64 (eqz)))
y = vbslq_f64 (eqz, x, y);
if (__glibc_unlikely (v_any_u64 (cmp))) if (__glibc_unlikely (v_any_u64 (cmp)))
return special_case (x, y, odd, cmp); return special_case (x, y, odd, cmp);
return vreinterpretq_f64_u64 (veorq_u64 (vreinterpretq_u64_f64 (y), odd)); return vreinterpretq_f64_u64 (veorq_u64 (vreinterpretq_u64_f64 (y), odd));

7
sysdeps/aarch64/fpu/sinf_advsimd.c
View File

@ -56,7 +56,7 @@ float32x4_t VPCS_ATTR V_NAME_F1 (sin) (float32x4_t x)
{ {
const struct data *d = ptr_barrier (&data); const struct data *d = ptr_barrier (&data);
float32x4_t n, r, r2, y; float32x4_t n, r, r2, y;
uint32x4_t odd, cmp, eqz; uint32x4_t odd, cmp;
#if WANT_SIMD_EXCEPT #if WANT_SIMD_EXCEPT
uint32x4_t ir = vreinterpretq_u32_f32 (vabsq_f32 (x)); uint32x4_t ir = vreinterpretq_u32_f32 (vabsq_f32 (x));
@ -70,7 +70,6 @@ float32x4_t VPCS_ATTR V_NAME_F1 (sin) (float32x4_t x)
cmp = vcageq_f32 (d->range_val, x); cmp = vcageq_f32 (d->range_val, x);
cmp = vceqzq_u32 (cmp); /* cmp = ~cmp. */ cmp = vceqzq_u32 (cmp); /* cmp = ~cmp. */
#endif #endif
eqz = vceqzq_f32 (x);
/* n = rint(|x|/pi) */ /* n = rint(|x|/pi) */
n = vfmaq_f32 (d->shift, d->inv_pi, r); n = vfmaq_f32 (d->shift, d->inv_pi, r);
@ -89,10 +88,6 @@ float32x4_t VPCS_ATTR V_NAME_F1 (sin) (float32x4_t x)
y = vfmaq_f32 (C (0), y, r2); y = vfmaq_f32 (C (0), y, r2);
y = vfmaq_f32 (r, vmulq_f32 (y, r2), r); y = vfmaq_f32 (r, vmulq_f32 (y, r2), r);
/* Sign of 0 is discarded by polynomial, so copy it back here. */
if (__glibc_unlikely (v_any_u32 (eqz)))
y = vbslq_f32 (eqz, x, y);
if (__glibc_unlikely (v_any_u32 (cmp))) if (__glibc_unlikely (v_any_u32 (cmp)))
return special_case (x, y, odd, cmp); return special_case (x, y, odd, cmp);
return vreinterpretq_f32_u32 (veorq_u32 (vreinterpretq_u32_f32 (y), odd)); return vreinterpretq_f32_u32 (veorq_u32 (vreinterpretq_u32_f32 (y), odd));