Showing posts with label float. Show all posts
Showing posts with label float. Show all posts

December 31, 2012

Arithmetic encoding (bookmark)

Based on the previous post, I can't stop until I implement my own arithmetic coding. The worse is I don't have my laptop with me, therefore I only have read articles and tried encoding/decoding on the paper. Which is slightly harder if numbers used are fixed point and want to imitate how C program should work. These 3 pages helped me most to understand the problem.

Arithmetic Code [michael.dipperstein.com]
Basic arithmetic coding [arturocampos.com]
Arithmetic coding how to [pdf] [webspace.ulbsibiu.ro/macarie.breazu]

April 15, 2012

Properties of the compiler and floating point arithmetic test

While i was reading The C library book, i come across two tests which are mentioned in the float chapter.

Enquire tests the properties of a compiler and the machine which is run on. It creates the file with the outputs.

Enquire test written by Steven Pemberton of CWI
More reading at: Enquire details

Paranoia tests the floating point arithmetic used.
Paranoia W.M. Kahan