If, after decimal point alignment, the absolute value of a result of an arithmetic operation exceeds the largest value that can be contained in the associated resultant-identifier, a size error condition exists. Division by zero always causes a size error condition. Note that the results of division by zero when no ON SIZE ERROR phrase is specified are unpredictable. The size error condition applies only to the final results, except in MULTIPLY and DIVIDE statements, in which case the size error condition applies to the intermediate results as well.
Violation of the rules for the evaluation of exponentiation always terminates the arithmetic operation and always causes a size error condition.
If the ROUNDED phrase is specified, rounding takes place before checking for size error. When such a size error condition occurs, the subsequent action depends on whether or not the SIZE ERROR phrase is specified, as follows: