Flash Memory LLR Update for Soft Decoding Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flash memory degradation makes it difficult to determine accurate soft decoding parameters, such as log-likelihood ratios (LLRs), which are crucial for error correction in data storage, leading to sub-optimal quantization and increased read latency.
Innovation Solution
A method that regularly updates LLRs using a blind technique, where the output of the ECC decoder is used as feedback to inform the LLRs, allowing for dynamic adaptation of quantization bins and parameters, even in the absence of known original inputs, thereby maintaining precision and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a larger number of reads are performed to improve precision, then the precision of unquantised output is improved, but read latency increases
Solution Approach 1:
The patent implements dynamic adjustment of read precision based on flash memory degradation state. The system transitions from static quantization to adaptive quantization where the number of read references is dynamically adjusted according to the degradation level detected through threshold voltage distribution analysis. This resolves the contradiction by allowing high precision only when necessary (when degradation is detected) while maintaining lower precision during normal operation.
Solution Approach 2:
The patent changes the parameter of read precision (number of read references) based on the degradation state of flash memory. By monitoring threshold voltage distributions and detecting shifts in charge states, the system adjusts the quantization parameter dynamically. This allows the system to maintain optimal precision for error correction without consistently performing excessive reads, thereby reducing latency while preserving accuracy when needed.
2Measurement precision
If soft decoding parameters are updated frequently to maintain accuracy, then decoding precision is improved, but processing overhead increases
Solution Approach 1:
The patent implements periodic updates of soft decoding parameters (LLRs) based on detected degradation thresholds rather than continuous updates. The system monitors flash memory state and triggers parameter updates only when degradation exceeds predetermined thresholds, creating a periodic rather than continuous update cycle. This reduces processing overhead while maintaining decoding accuracy by updating parameters only when necessary.
Solution Approach 2:
The patent employs feedback mechanisms where the output of the ECC decoder and observed error patterns are used to inform and adjust LLR values. The system uses the decoded output and error statistics as feedback to refine soft decoding parameters, creating a closed-loop system that automatically adapts to degradation without requiring external intervention or excessive processing of reference data.
3Quantity of substance
If flash memory is used extensively to provide storage capacity, then storage capacity is improved, but reliability of soft decoding parameters deteriorates due to degradation
Solution Approach 1:
The patent performs preliminary characterization of flash memory threshold voltage distributions during manufacturing or initial use to establish baseline LLR values and degradation thresholds. By pre-characterizing the memory device and storing reference data about its specific properties, the system is prepared to detect degradation early and adjust parameters accordingly, maintaining reliability even as the memory is extensively used for storage.
Solution Approach 2:
The patent implements self-service mechanisms where the flash memory system automatically monitors its own degradation state and adjusts its soft decoding parameters without external intervention. The system uses its own decoded output and error patterns as feedback to self-correct and maintain parameter accuracy, enabling reliable operation throughout the storage device's lifecycle without requiring manual recalibration or external testing.
Data Source
AI summary
A method performed in a computing device. The computing device is configured to store data and retrieve stored data from storage. The computing device further stores parameters for use in soft decoding stored data. The method comprises retrieving data from storage using soft decoding based on the stored soft decoding parameters, using retrieved, soft decoded data to estimate updates of one or more of the parameters for soft decoding and storing the updates.