Convert packed unsigned 16-bit integers in a to packed single-precision (32-bit) floating-point elements.
See Implementation
Convert packed unsigned 16-bit integers in a to packed single-precision (32-bit) floating-point elements.