[__arm_]vcvttq_f32_f16

Convert between half-precision and single-precision floating-point values in elements of a vector register. For half-precision to single-precision operation, the top half (T variant) or bottom half (B variant) of the source vector register is selected. For single-precision to half-precision operation, the top half (T variant) or bottom half (B variant) of the destination vector register is selected and the other half retains its previous value.