Floating point

If you’ve ever worked with compilers, and floating point numbers — you’ll know sometimes you get very odd results…

This is by and large due to the way floating point works, and how IEEE- compliancy is assured. This excellent paper explains why and how all those things happen, as well as being an excellent grounding in all things floating point.

When you work on systems with retarded FPUs (read: ps2) this kind of thing is nice to know, even if its actual usage is academic!

Ok, not strictly C++ specific, but still interesting, and of relevance if you do any kind of work with floating point numbers.

Filed under: Coding
Posted at 17:22:50 BST on 14th June 2004.

About Matt Godbolt

Matt Godbolt is a C++ developer working in Chicago in the finance industry.