Calculations involving floating point values often produce results that are not what you expect. If you perform a floating point calculation and then compare the results against some expected value, it is unlikely that you get the intended result.
Instead of comparing the results of a floating point calculation, make sure that the result is greater or less than what is needed, with a given error. Here is an example:
SELECT i, SUM(a) as sum_a, SUM(b) as sum_b FROM t1 GROUP BY i HAVING ABS(sum_a - sum_b) > 1E-10;