user0/rust
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

convert _mm256_i64gather_pd to const generics

Browse source
This commit is contained in:
Rémy Rakic 2021-03-06 01:47:46 +01:00 • committed by Amanieu d'Antras
parent 6e89c87395
commit b0fcace623
1 changed files with 9 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

19
library/stdarch/crates/core_arch/src/x86/avx2.rs
View file

@ -1774,20 +1774,19 @@ pub unsafe fn _mm_mask_i64gather_pd<const SCALE: i32>(
/// [Intel's documentation](https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm256_i64gather_pd)
#[inline]
#[target_feature(enable = "avx2")]
#[cfg_attr(test, assert_instr(vgatherqpd, scale = 1))]
#[rustc_args_required_const(2)]
#[cfg_attr(test, assert_instr(vgatherqpd, SCALE = 1))]
#[rustc_legacy_const_generics(2)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm256_i64gather_pd(slice: *const f64, offsets: __m256i, scale: i32) -> __m256d {
pub unsafe fn _mm256_i64gather_pd<const SCALE: i32>(
slice: *const f64,
offsets: __m256i,
) -> __m256d {
static_assert_imm8_scale!(SCALE);
let zero = _mm256_setzero_pd();
let neg_one = _mm256_set1_pd(-1.0);
let slice = slice as *const i8;
let offsets = offsets.as_i64x4();
macro_rules! call {
($imm8:expr) => {
vpgatherqpd(zero, slice, offsets, neg_one, $imm8)
};
}
constify_imm8_gather!(scale, call)
vpgatherqpd(zero, slice, offsets, neg_one, SCALE as i8)
}
/// Returns values from `slice` at offsets determined by `offsets * scale`,
@ -5905,7 +5904,7 @@ mod tests {
j += 1.0;
}
// A multiplier of 8 is word-addressing for f64s
let r = _mm256_i64gather_pd(arr.as_ptr(), _mm256_setr_epi64x(0, 16, 32, 48), 8);
let r = _mm256_i64gather_pd::<8>(arr.as_ptr(), _mm256_setr_epi64x(0, 16, 32, 48));
assert_eq_m256d(r, _mm256_setr_pd(0.0, 16.0, 32.0, 48.0));
}