Function f32x4_relaxed_nmadd

#[target_feature(enable = "relaxed-simd")]
pub fn f32x4_relaxed_nmadd(a: v128, b: v128, c: v128) -> v128

Computes -a * b + c with either one rounding or two roundings.