Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements.
See Implementation
Convert packed signed 32-bit integers in a to packed double-precision (64-bit) floating-point elements.