rptu.de/channel-codes: An Update on the Maximum Likelihood Decoding Performance of 5G-NR Channel Codes
Oliver Griebel, Kira Kraft, Lucas Johannsen, Claus Kestel, Gabriel Machado Dick, Norbert Wehn · 2025
Maximum likelihood (ML) decoding is the "champions league" of all decoding algorithms. Instead of solving the decoding problem heuristically, it is solved optimally and, therefore, provides the best possible error correction performance among all decoding algorithms. On the downside, ML decoding is an NP-hard problem. Knowing the ML performance of different codes is important as it provides an ultimate benchmark for the code’s potential and the potential of any heuristic decoding algorithm, which is especially valuable for the research community. This paper provides the state of the art on simulating a code’s ML performance and new results for 5G new radio (5G-NR) low-density parity-check (LDPC) and polar codes. All results are publicly available at www.rptu.de/channel-codes, which is well known by the coding community.