According to large-page-size and random-bit-error characteristics, long-block-length Bose-Chaudhuri-Hochquenghem (BCH) decoders are applied to realize error correction in nand Flash memory devices. To accelerate the decoding process in an area-efficient architecture, a parallel architecture with minimal polynomial combinational network (MPCN) for long BCH decoders is presented in this brief. The proposed design utilizes MPCNs to replace constant finite-field multipliers, which dominate the hardware complexity of the high-parallel Chien search architecture. Furthermore, both the syndrome calculator and the Chien search can be merged by exploiting our MPCN-based architecture, leading to significant hardware complexity reduction. From the synthesis results in the 90-nm CMOS technology, the MPCN-based joint syndrome calculation and Chien search has 46.7% gate count saving for parallel-32 BCH (4603, 4096; 39) decoder in contrast with the straightforward design.
|頁（從 - 到）||682-686|
|期刊||IEEE Transactions on Circuits and Systems I: Regular Papers|
|出版狀態||Published - 1 10月 2011|