Getting around the rounding error in binary division

[Hilo890]

How do x86 processors usually go about dividing numbers while getting around the rounding error of binary division? That's probably a broad question with a lot of answers, but if you know any techniques they use to divide numbers, I'd like to know.

[Duthomhas]

They don't -- the rounding error propagates.

[Hilo890]

But how do computers deal with it? Why don't we just have numbers really close to the actual answer when you do division on a computer in whatever program? Rounding? But then if the answer was actually supposed to be that number non-rounded, it would be wrong... Clearly I'm pretty ignorant on the subject. Haha.

[Duthomhas]

Computers don't deal with it -- they are too stupid.

A smart programmer will organize his mathematical expressions in a way to avoid as much error as he needs.

It might be worth your time to read through this:
http://perso.ens-lyon.fr/jean-michel.muller/goldberg.pdf