I think what Jay Box was refering to was floating and fixed-point maths, that is quite complicated to implement into any controller, let alone an 8-bit one!! If you are just dividing simple integer signed/unsigned 8-bit or 16-bit numbers and only require either a rounded result, or a result and remainder, the routines are usally relativley simple (clever, but simple!)

Atmel seem to have some good application notes covering maths routines for their AVR micro-controllers, even if you're not familiar with AVR op-codes, it should't be difficult to work out whats going on in their code. I use 8052s and AVRs and have never found it a problem switching between working with the two.

Have a look at their app. note: AVR200

It can be found at:
http://www.atmel.com/atmel/products/prod201.htm

Hope I've helped