Design of high-throughput re-encoder for soft-decision Reed-Solomon decoding

Abstract

The re-encoding and coordinate transformation techniques are usually employed to reduce the computational complexity of interpolation, which is the most computation-intensive step in the algebraic soft-decision decoding of Reed-Solomon (RS) codes. In this paper, we present a high-throughput re-encoder design for soft-decision decoding a (255, 239) RS code. With the developed scheduling scheme and a folding architecture, the resulting design can significantly shorten the latency of the re-encoding process to achieve a high throughput rate and effectively reduce the hardware requirement of the re-encoder. Compared to related work, the developed re-encoder has a 4-fold increase in the throughput rate with only 16% hardware overhead.