Comparison of different schedulings for the ADMM based LDPC decoding

Horizontal Layered Scheduling ADMM Penalized Decoder Based on the Improved Penalty Function for LDPC

“…Based on this discovery, to reduce the ADMM decod er's complexity, Wei et al designed a method by reducing the number of Euclidean projections rather than simplifying the Euclidean projection operation [11]. In terms of the ADMM decoder, the HL scheduling [12] and the VL scheduling [13] can speed up decoding compared with the flooding scheduling. This result occurs because the iterations needed to converge to a valid codeword is much lower for the layered scheduling wh i ch le a d s to a mea n in g f u l co m p u t at i o n sav i n g compared with the flooding scheduling.…”

“…First, the maximum computational complexity and memory complexity of the flooding scheduling, HL scheduling and VL scheduling ADMM penalized decoders are compared in Table 1 [13 ]. q represents the number of decoding iterations, ( )…”

“…Choosing a suitable scheduling strategy for the ADMM decoder is another way to increase its convergence speed. For example, the horizontal layered (HL) scheduling [12], vertical layered (VL) scheduling [13], and the node-wise scheduling [14] can achieve the above result to varying degrees. By adopting the approximate line segment projection algorithm, Xia et al designed the modified node-wise scheduling algorithm that can simplify the node-wise scheduling ADMM decoding method [15].…”

Novel Early Termination Method of an ADMM-Penalized Decoder for LDPC Codes in the IoT

“…Another way to reduce the ADMM-based LP decoder's complexity is the message scheduling strategy. For the ADMM decoding method, horizontal-layered scheduling [29], verticallayered scheduling [30], and node-wise scheduling [31] are presented to increase the convergence rate.…”