What Every Computer Scientist Should Know About Floating-Point Arithmetic

Floating point numbers always seem to confuse programmers, including myself, that haven’t fully looked into the IEEE standard that dictates how they are stored and manipulated in most modern languages.  My System Architecture professor recommended this article to out class to help us understand a portion of the MIPS architecture.

It is somewhat lengthy at 72 pages, but I can honestly say it is one of the most informative papers I have ever read.  Literally everything you have ever wanted to know about floating point standards is covered.  It covers everything from how the numbers are stored at the binary level all the way through how mathematical operations are performed on them.

Really, it’s worth reading.  Even if you don’t read all of the mathematical proofs at the end, at least glance at the IEEE standard way of storing floating-point numbers and the section on accuracy.  I guarantee it will make the whole topic far easier to grasp.

Link to article: http://docs.sun.com/source/806-3568/ncg_goldberg.html