/EXTENDED_FLOAT
       /X_FLOAT

    (Alpha only) Converts the expression on the right-hand side of
    the equal sign to the IEEE X_floating type (length 16 bytes) and
    deposits the result into the specified location.