Adaptive LDPC Decoding Across Failed Memory Word Lines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In memory systems, especially SSDs, soft information generated during LDPC decoding is not effectively utilized after initial error correction, leading to suboptimal reliability adjustments based on iteration and unsatisfied check information, particularly in scenarios with multiple failed word lines.
Innovation Solution
Implementing a method where the memory controller performs soft decoding across all word lines, generates extrinsic information for failed word lines, updates soft information based on this extrinsic information, and propagates it across these lines, using scaling factors adjusted by unsatisfied check and iteration data to enhance LDPC decoding reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If soft decoding is performed on all word lines to generate codewords, then error correction capability is improved, but decoding complexity increases
Solution Approach 1:
The decoding process is segmented into two phases: first, soft decoding is performed on all word lines to generate initial codewords and soft information; second, for failed word lines only, extrinsic information is generated and soft information is updated and propagated. This segmentation allows the system to maintain high error correction capability while reducing overall decoding complexity by applying intensive processing only where needed.
Solution Approach 2:
Soft decoding is performed preliminarily on all word lines before final error correction. This preliminary action generates soft information that can be reused and propagated to failed word lines, avoiding the need to perform full soft decoding on every word line and thus reducing complexity while maintaining reliability.
2Reliability
If extrinsic information is generated and soft information is updated for failed word lines, then decoding reliability is improved, but processing time increases
Solution Approach 1:
The intensive extrinsic information generation and soft information update operations are applied locally only to failed word lines rather than uniformly to all word lines. This local quality approach ensures high decoding reliability for failed lines while minimizing the overall processing time by avoiding redundant operations on successfully decoded lines.
Solution Approach 2:
The system uses feedback from the initial soft decoding results to identify which word lines failed and require additional processing. Extrinsic information is generated based on this feedback, and soft information is propagated selectively to failed word lines, optimizing the balance between reliability improvement and processing time.
3Reliability
If soft information is propagated across failed word lines, then data recovery capability is improved, but computational load increases
Solution Approach 1:
The soft information propagation is performed dynamically based on the decoding status of each word line. The system adapts the propagation process to the specific needs of failed word lines, adjusting the computational load according to the actual data recovery requirements rather than applying a fixed high-complexity process to all lines.
Data Source
AI summary
Systems, memory controllers, decoders and methods perform decoding by exploiting differences among word lines for which soft decoding fails (failed word lines). Such decoding generates extrinsic information for codewords of failed word lines based on the soft decoding. The soft information obtained during the soft decoding is updated based on the extrinsic information, and the updated soft information is propagated across failed word lines. Low-density parity-check (LDPC) decoding of codewords of failed word lines is performed with the updated soft information.