Function vrndx_f16

#[target_feature(enable = "neon", enable = "fp16")]
pub fn vrndx_f16(a: float16x4_t) -> float16x4_t

Floating-point round to integral exact, using current rounding mode Arm's documentation