Memory Soft-Decision Compression Using Minor-Bit Position Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory devices face challenges in efficiently processing soft decision data for error correction due to the inefficiencies in data compression and the high probability of error correction failure when threshold voltage distributions overlap, leading to uncorrectable errors.
Innovation Solution
A memory device that performs fixed-size compression on soft decision data by encoding the position of minor bits and hierarchically compressing soft decision data into segments, using a compression circuit to generate compressed data, and a decompression circuit to restore the data for error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If soft decision data is compressed by encoding position values of minor bits, then data size is reduced, but some bit information is lost
Solution Approach 1:
The patent extracts only the position values of minor bits (bits with first value) from the soft decision data, discarding the redundant information about bits with the other value. This extraction approach reduces data size while preserving the essential reliability information needed for error correction, as the position of minor bits contains the critical reliability differentiation.
Solution Approach 2:
The patent changes the representation parameter of soft decision data from storing actual bit values to storing position values of minor bits. This parameter transformation enables fixed-size compression segments to be generated, reducing the overall data size while maintaining the reliability information necessary for effective error correction operations.
2Productivity
If fixed size compression is applied to soft decision data, then data processing efficiency is improved, but data structure complexity increases
Solution Approach 1:
The patent divides soft decision data into multiple fixed-size compression segments, where each segment contains position values of minor bits from a corresponding sub-segment. This segmentation approach standardizes the data structure, enabling more efficient processing and storage while the regular fixed-size format actually simplifies the overall data structure compared to variable-length representations.
Solution Approach 2:
The patent creates a simplified copy of the essential reliability information by representing soft decision data as position values rather than full bit sequences. This copied representation maintains the critical information needed for error correction while reducing data size and simplifying the data structure for more efficient processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A memory device includes first and second memory cell arrays each including a plurality of memory cells, a page buffer circuit configured to read first soft decision data including a plurality of sub-segments from the first memory cell array, and a compression circuit configured to perform a first compression operation of generating a first compression segment including a number of position values less than or equal to a first reference number, on each of a plurality of partial segments included in one of the plurality of sub-segments and sequentially perform a plurality of compression operations, which is subsequent to the first compression operation, of generating a next compression segment including a number of position values, which are less than or equal to a reference number corresponding to each compression operation, of position values included in two or more previous compression segments in each compression operation.