compiler_builtins::math::libm_math::generic::fma_wide

Function fma_wide_round

Source 
pub fn fma_wide_round<F, B>(x: F, y: F, z: F, round: Round) -> FpResult<F>
where
    F: Float + HFloat<D = B>,
    B: Float + DFloat<H = F>,
    B::Int: CastInto<i32>,
    i32: CastFrom<i32>,
Expand description

Fma implementation when a hardware-backed larger float type is available. For f32 and f64, f64 has enough precision to represent the f32 in its entirety, except for double rounding.