Computational complexity and quantization effects of decoding algorithms for non-binary LDPC codes

28 September 2004

conference paper
Published by Institute of Electrical and Electronics Engineers (IEEE)

Vol. 4 (15206149)
https://doi.org/10.1109/icassp.2004.1326915

Abstract

This contribution deals with the comparison of the sum-product algorithm (SPA) and its log-domain version (log-SPA) for decoding LDPC (low density parity check) codes over general binary extension fields. For both algorithms, we determine their computational complexity based on the number of real-valued operations and investigate their sensitivity to quantization effects. Whereas the log-SPA yields the shorter decoding time in the case of binary LDPC codes, we point out that increasing the field size tends to favor the SPA, especially when a multiplication takes only slightly more time than an addition. Further, we show that log-SPA requires fewer quantization levels and suffers less from a quantization induced error-floor.

Keywords

This publication has 9 references indexed in Scilit:

On finite precision implementation of low density parity check codes decoder
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
Design of fixed-point iterative decoders for concatenated codes with interleavers
IEEE Journal on Selected Areas in Communications, 2001
Design of capacity-approaching irregular low-density parity-check codes
IEEE Transactions on Information Theory, 2001
Factor graphs and the sum-product algorithm
IEEE Transactions on Information Theory, 2001
Decoding low density parity check codes with finite quantization bits
IEEE Communications Letters, 2000
Decoding and Equalization with Analog Non‐linear Networks
European Transactions on Telecommunications, 1999
Good error-correcting codes based on very sparse matrices
IEEE Transactions on Information Theory, 1999
Low-density parity check codes over GF(q)
IEEE Communications Letters, 1998
Low-density parity-check codes
IEEE Transactions on Information Theory, 1962