Encyclopedia of Sparse Graph Codes

David J C MacKay

Evaluation of Gallager codes for low error tolerance, short block length and high rate applications. Sparse graph codes include Gallager codes, Tanner codes, MN codes Repeat-Accumulate codes (RA codes), and turbo codes, all of which have near-Shannon limit performance. This paper (which is still in preparation) describes empirical properties of a wide selection of these codes, comparing in particular the codes' block error rates with an emphasis on undetected versus detected errors. We explore the dependence of block error rate on block length and other code construction parameters. Histograms of decoding time are also shown. Draft 1 concentrates on small block lengths.

postscript (Cambridge UK).

postscript (Canada mirror).

@unpublished{MacKayEncyclopedia98,
 KEY            ="",
 AUTHOR         ="D. J. C.  MacKay",
 TITLE          ="Encyclopedia of Sparse Graph Codes", 
 ADDRESS	="",
 YEAR           ="1998",
 PAGES		="",
note={{\verb|http://www.inference.phy.cam.ac.uk/mackay/|}},
 ANNOTE ="Date submitted: ; Date accepted: ;  MRAO "}
@unpublished{MacKay99ENC,
 KEY            ="",
 AUTHOR         ="D. J. C.  MacKay",
 TITLE          ="Encyclopedia of Sparse Graph Codes (hypertext archive)", 
 ADDRESS	="",
 YEAR           ="1999",
 PAGES		="",
note={{\verb|http://www.inference.phy.cam.ac.uk/mackay/codes/data.html|}},
 ANNOTE ="Date submitted: ; Date accepted: ;  MRAO "}

David MacKay's: home page, publications. bibtex file.

Canadian mirrors: home page, publications. bibtex file.