David MacKay's Gallager code resources

Contents:

• Papers and books
• Pictorial demonstration of iterative decoding
• Matrices used for those demonstrations
• Database containing performance results for many sparse graph codes; also includes some parity check matrices
• Small archive of codes over GF(4) and GF(8)
• Comparison of Performance of Codes for Quantum Error Correction
• Tutorial documents on Digital fountain codes
• Source code for modular encoding, transmission and sum-product decoding (by David MacKay and Radford Neal)

• Source code for lots of code-related functions: including "code6" for making Gallager codes; "mncN" and "gallager" for sum-product decoding and "fe" for decoding using variational free energy minimization (not the best decoder; sum-product is better) Before making executables, you may need to create some directories in both . and ./ansi thus:

   mkdir bin$ARC; cd bin$ARC ; ln -s ../_Makefile Makefile ; cd .. 
   OR for example:
   mkdir bini386; cd bini386; ln -s ../_Makefile Makefile ; cd .. 

  [The main programs I use in my work on Gallager codes are code6 and gallager.]

  The software is provided under the GNU General Public License.

• Digital fountain demonstration (LT codes) (octave)
• Raw results on many codes
• Documentation for perl program `GH.p' which may be used to generate `alist matrices'
• "spec" files for GHG.p
• Source code for Progressive Edge Growth parity-check matrix construction, supplied by Xiao-Yu Hu [xhu:-:AT:-:zurich.ibm.com]
• Capacity curves for the binary-input Gaussian channel, by Matt Davey.