Memory ECC Bit Flipping with Dynamic Soft-Bit Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory sub-systems face inefficiencies in error correction due to the susceptibility of bit-flipping algorithms to noise and lack of dynamic adjustment in threshold values for bit flipping, leading to suboptimal decoding efficiency and reliability.
Innovation Solution
A memory sub-system controller utilizes soft bits and match bits to dynamically adjust bit flip thresholds during error correcting code operations, using a threshold table to determine whether to flip bits based on the number of unsatisfied check nodes, thereby improving decoding efficiency and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fixed threshold values are used for bit flipping in error correction, then the decoding process is simple and fast, but the reliability and decoding efficiency deteriorate due to susceptibility to noise and lack of dynamic adjustment
Solution Approach 1:
The patent applies dynamics by transitioning from fixed threshold values to dynamic threshold values that adapt during the decoding process. The threshold is adjusted based on the iteration number and the number of unsatisfied check nodes, allowing the system to respond to changing conditions and improve decoding reliability while maintaining manageable complexity through structured adjustment rules.
Solution Approach 2:
The patent implements parameter changes by modifying the threshold parameter based on decoding progress. Specifically, the threshold value changes as a function of the iteration number and the count of unsatisfied check nodes, enabling the error correction algorithm to adapt its behavior to the current state of decoding and thereby improve reliability without excessive complexity.
2Productivity
If dynamic bit flip thresholds are implemented based on soft bit and match bit, then decoding efficiency and reliability improve, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies preliminary action by pre-defining the threshold adjustment strategy based on iteration number and unsatisfied check node count. This structured approach allows the system to make efficient decisions during decoding without requiring complex real-time calculations, thereby improving decoding efficiency while limiting the additional processing time to predetermined adjustments.
Solution Approach 2:
The patent implements feedback by using the number of unsatisfied check nodes as a metric to dynamically adjust the threshold. This feedback mechanism allows the decoding process to adapt to the current error conditions, improving efficiency by focusing corrections where needed while the structured feedback loop prevents excessive processing time through clear termination criteria.
3Reliability
If traditional error correction methods are used without dynamic threshold adjustment, then the processing is faster and simpler, but the performance and longevity of the memory sub-system decrease
Solution Approach 1:
The patent applies dynamics by implementing a memory sub-system that adapts its error correction behavior through dynamic threshold adjustment. The threshold evolves during decoding based on performance metrics, enabling the system to improve reliability and longevity while maintaining acceptable complexity through structured adaptation rules rather than arbitrary complexity.
Data Source
AI summary
A soft input is obtained from a sense word corresponding to encoded host data read from the memory device and decoded using a parity-check matrix. A match array is maintained. Each iteration of an error correcting code operation a number of unsatisfied check nodes of a respective bit of the sense word is calculated for each bit of the sense word. A bit flip threshold from a threshold data structure is obtained based on a current iteration of the error correcting code operation, a soft bit associated with the respective bit, and a match bit associated with the respective bit. The respective bit is flipped based on the number of unsatisfied check nodes satisfying the bit flip threshold.