Function vrndnq_f16

#[target_feature(enable = "neon", enable = "fp16")]
pub fn vrndnq_f16(a: float16x8_t) -> float16x8_t

Floating-point round to integral, to nearest with ties to even Arm's documentation