Internal report (Reports)
Early-Stopped Approach and Analysis for the Berlekamp-Massey Algorithm
Chou, Hung-Pu; Hong-fu, Chou
2022
 

Files


Full Text
A_Novel_Early_Stopped_Approach_and_Analysis_for_the_Berlekamp_Massey_Algorithm.pdf
Publisher postprint (586.17 kB)
Download

All documents in ORBilu are protected by a user license.

Send to



Details



Keywords :
BCH code; BM algorithm; low latency design
Abstract :
[en] BCH codes are being widely used in commercial NAND flash controllers, and the decoding algorithm based on the Berlekamp-Massey (BM) algorithm is a classic solution for solving the key equation used for error correction. The latency of BM decoding is the bottleneck of the Bose-Chaudhuri Hocquenghem (BCH) decoder when correcting a high number of bit errors. However, the flash memory has an error distribution that degrades with usage: few errors occur in the new memory and a low number of errors occur within a code block. With usage, the system performance degrades and BM decoding needs t iterations in order to correct a larger number t of errors. In an attempt to improve the system performance for high speed applications, early termination of the BM decoding is necessary to overcome this degradation. In this paper, a practical solution for early termination checking for BM algorithm is provided. The analysis of proposed method is presented by means of considering the weight distribution of BCH code and deriving the probability of malfunction as the event of undetectable error. The proposed method is presented to be effective by the numerical results and the probability of malfunction for the proposed method is lower than 10−26. As a result, the FPGA testing on a USB device validate the reliability of the proposed method for applying to a commercial product.
Disciplines :
Computer science
Author, co-author :
Chou, Hung-Pu ;  University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > SigCom
Hong-fu, Chou
Language :
English
Title :
Early-Stopped Approach and Analysis for the Berlekamp-Massey Algorithm
Publication date :
2022
Publisher :
Luxembourg University, Luxembourg, Luxembourg
Focus Area :
Computational Sciences
Available on ORBilu :
since 13 September 2022

Statistics


Number of views
68 (21 by Unilu)
Number of downloads
73 (6 by Unilu)

Bibliography


Similar publications



Contact ORBilu