ISNORMAL returns a '1'B if if the argument with which it
is invoked is not a zero, subnormal, infinity or NaN. Otherwise it
returns a '0'B.
>>-ISNORMAL(x)-------------------------------------------------><
- x
- REAL FLOAT DECIMAL expression.
The FLOAT(DFP) compiler option must be in effect.
No floating-point exceptions will be raised no matter what the format of the argument.