Convert a double-precision floating-point number to the nearest single-precision floating-point number.
