1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51

//! Vector fused multiply add
#![allow(dead_code)]
use coresimd::simd::*;

#[allow(improper_ctypes)]
extern "C" {
    #[link_name = "llvm.fma.v2f32"]
    fn fma_v2f32(x: f32x2, y: f32x2, z: f32x2) -> f32x2;
    #[link_name = "llvm.fma.v4f32"]
    fn fma_v4f32(x: f32x4, y: f32x4, z: f32x4) -> f32x4;
    #[link_name = "llvm.fma.v8f32"]
    fn fma_v8f32(x: f32x8, y: f32x8, z: f32x8) -> f32x8;
    #[link_name = "llvm.fma.v16f32"]
    fn fma_v16f32(x: f32x16, y: f32x16, z: f32x16) -> f32x16;
    #[link_name = "llvm.fma.v2f64"]
    fn fma_v2f64(x: f64x2, y: f64x2, z: f64x2) -> f64x2;
    #[link_name = "llvm.fma.v4f64"]
    fn fma_v4f64(x: f64x4, y: f64x4, z: f64x4) -> f64x4;
    #[link_name = "llvm.fma.v8f64"]
    fn fma_v8f64(x: f64x8, y: f64x8, z: f64x8) -> f64x8;
}

pub(crate) trait FloatFma {
    fn fma(self, y: Self, z: Self) -> Self;
}

macro_rules! impl_fma {
    ($id:ident : $fn:ident) => {
        #[cfg(not(target_arch = "s390x"))]
        impl FloatFma for $id {
            fn fma(self, y: Self, z: Self) -> Self {
                unsafe { $fn(self, y, z) }
            }
        }
        // FIXME: https://github.com/rust-lang-nursery/stdsimd/issues/501
        #[cfg(target_arch = "s390x")]
        impl FloatFma for $id {
            fn fma(self, y: Self, z: Self) -> Self {
                self * y + z
            }
        }
    };
}

impl_fma!(f32x2: fma_v2f32);
impl_fma!(f32x4: fma_v4f32);
impl_fma!(f32x8: fma_v8f32);
impl_fma!(f32x16: fma_v16f32);
impl_fma!(f64x2: fma_v2f64);
impl_fma!(f64x4: fma_v4f64);
impl_fma!(f64x8: fma_v8f64);