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

sse3: _mm_movehdup_ps

Browse source
This commit is contained in:
gwenn 2017-09-30 13:11:08 +02:00 • committed by Andrew Gallant
parent 8e07404403
commit 261534cb0f
1 changed files with 17 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

18
library/stdarch/src/x86/sse3.rs
View file

@ -1,5 +1,5 @@
use x86::__m128i;
use simd_llvm::simd_shuffle2;
use simd_llvm::{simd_shuffle2, simd_shuffle4};
use v128::*;
#[cfg(test)]
@ -78,6 +78,15 @@ pub unsafe fn _mm_movedup_pd(a: f64x2) -> f64x2 {
simd_shuffle2(a, a, [0, 0])
}
/// Duplicate odd-indexed single-precision (32-bit) floating-point elements
/// from `a`.
#[inline(always)]
#[target_feature = "+sse3"]
#[cfg_attr(test, assert_instr(movshdup))]
pub unsafe fn _mm_movehdup_ps(a: f32x4) -> f32x4 {
simd_shuffle4(a, a, [1, 1, 3, 3])
}
#[allow(improper_ctypes)]
extern {
#[link_name = "llvm.x86.sse3.addsub.ps"]
@ -165,4 +174,11 @@ mod tests {
let r = sse3::_mm_movedup_pd(a);
assert_eq!(r, f64x2::new(-1.0, -1.0));
}
#[simd_test = "sse3"]
unsafe fn _mm_movehdup_ps() {
let a = f32x4::new(-1.0, 5.0, 0.0, -10.0);
let r = sse3::_mm_movehdup_ps(a);
assert_eq!(r, f32x4::new(5.0, 5.0, -10.0, -10.0));
}
}