I agree with Sheloo for the OP's question as stated. Integers up to
+/-10^15 can be exactly represented in floating point, and the result
should (and does) sum precisely to zero if manually entered (as opposed
to being the result of a formula that could have rounding errors).


--
shg
------------------------------------------------------------------------
shg's Profile: http://www.thecodecage.com/forumz/member.php?userid=13
View this thread: http://www.thecodecage.com/forumz/sh...ad.php?t=24483