Computational Complexity Analysis of FEC Decoding on SDR Platforms

The computational complexity evaluation is necessary for software defined Forward Error Correction (FEC) decoders. However, currently there are a limited number of literatures concerning on the FEC complexity evaluation using analytical methods. In this paper, three high efficient coding schemes including Turbo, QC-LDPC and Convolutional code (CC) are investigated. The hardware-friendly decoding pseudo-codes are provided with explicit parallel execution and memory access procedure. For each step of the pseudo-codes, the parallelism and the operations in each processing element are given. Based on it the total amount of operations is derived. The comparison of the decoding complexity among these FEC algorithms is presented, and the percentage of each computation step is illustrated. The requirements for attaining the evaluated results and reference hardware platforms are provided. The benchmarks of state-of-the-art SDR platforms are compared with the proposed evaluations. The analytical FEC complexity results are beneficial for the design and optimization of high throughput software defined FEC decoding platforms.