研究目的
To correct errors of read-hot data in 3D-TLC NAND flash by proposing a Read-disturb Modeled Artificial Neural Network Coupled LDPC ECC (RDNN-LDPC).
研究成果
The proposed RDNN-LDPC effectively corrects read-disturb errors in 3D-TLC NAND flash, extending acceptable read cycles by 10-times compared to conventional ANN-LDPC.
研究不足
The study focuses on read-hot data error correction in 3D-TLC NAND flash, potentially limiting applicability to other types of flash memory or error types.
1:Experimental Design and Method Selection:
The study proposes RDNN-LDPC to correct read-disturb errors in 3D-TLC NAND flash, analyzing how input parameters and model changes affect error correction.
2:Sample Selection and Data Sources:
Uses read-hot data from 3D-TLC NAND flash for training RDNN-LDPC.
3:List of Experimental Equipment and Materials:
3D-TLC NAND flash memory, SSD controller with RDNN-LDPC.
4:Experimental Procedures and Operational Workflow:
Training RDNN-LDPC with read-hot data, predicting BER, and calculating LLR for LDPC decoding.
5:Data Analysis Methods:
Comparison of predicted BER with measured BER, evaluation of acceptable read cycles extension.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容