Layered LDPC decoder in-order message access scheduling: a case study

This paper discusses the problems of message map-ping and message access scheduling for data hazards avoidance in layered Low-Density Parity-Check Decoder architectures, by using in-order message update strategy. We provide a detailed description of the offline algorithm that provides the fine tuning of message access scheduling that mitigates pipeline related hazards. An increase of 25%, up to 66% in hardware usage efficiency has been obtained when using these algorithms, for WiMAX rate 3/4 code, for the minimum latency case (2 clock cycles), with respect to the case when no off-line optimizations are used. Similar results have been obtained for DVB-S2X code, with 10%, up to 47%, showing that the fine-tune access scheduling improves the throughput of the LDPC decoder architecture at the same cost.