Successive-cancellation list decoder of polar codes based on GPU

In this paper, a Graphics Processing Unit (GPU) implementation of a Successive-Cancellation List (SCL) decoder for polar codes is proposed. Based on the compute unified device architecture (CUDA), the strategy to parallelize the decoding procedure is designed to reduce the latency. Moreover, the optimal design of data structures for several kinds of intermediate information is presented. In order to reduce the global memory accesses in list pruning, The GPU-adapted lazy-copy strategy is elaborated. By means of the parallel computing capabilities of GPUs, the proposed decoder achieves high throughput of 41Mbps on NVIDIA GTX 980 and 65Mbps on TITAN X while decoding the code with length of 1024 bits and 32 lists.