Memory Device Error Count Scaling for Compact Reliability Reporting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory devices face challenges in efficiently storing and reporting error counts due to the large storage space required for high-granularity metadata, which can grow significantly over the device's lifespan, necessitating a more efficient method to convey reliability and lifespan information with reduced storage requirements.
Innovation Solution
Implementing a scaling algorithm to convert raw error counts into scaled error counts using histogram bins, either linearly or logarithmically, and storing these scaled counts in user-accessible mode registers for reporting, allowing for reduced storage needs while maintaining valuable performance information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-granularity metadata is stored to provide detailed error count information, then measurement precision is improved, but storage space requirements increase significantly
Solution Approach 1:
The patent applies parameter changes by transforming the error count data through scaling operations. The raw error count is divided by a scaling factor and rounded to produce a scaled error count, which reduces the numerical value while preserving the relative magnitude relationships. This parameter transformation allows the same information to be represented with fewer bits, directly resolving the contradiction between precision and storage space.
Solution Approach 2:
The patent segments the error count information into multiple fields with different precision levels. The scaled error count field provides coarse-grained information for most cases, while the raw error count field is only stored and transmitted when necessary (e.g., when error counts exceed thresholds). This segmentation allows the system to maintain measurement precision when needed while minimizing storage space requirements for typical operations.
2Reliability
If all metadata is stored to provide complete reliability information, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the essential information from the complete metadata set. Instead of storing all raw error counts and detailed metadata, the system extracts and stores only the scaled error counts in user-accessible locations. This extraction process maintains the core reliability information while significantly reducing the complexity of metadata management and storage requirements.
Solution Approach 2:
The patent implements dynamic metadata management where the level of detail stored and accessed varies based on operational conditions. The system dynamically switches between storing scaled error counts for normal operations and raw error counts when thresholds are exceeded. This dynamic approach allows the system to maintain reliability information completeness while adapting the complexity of metadata management to actual needs.
3Measurement precision
If error counts are reported with high precision to enable accurate reliability assessment, then measurement precision is improved, but loss of information decreases (more information to manage)
Solution Approach 1:
The patent uses parameter changes through scaling to maintain the essential information content while reducing the magnitude of numbers. By dividing the raw error count by a scaling factor and rounding, the system preserves the relative magnitude relationships and trend information while reducing the information management burden. The scaled error count maintains measurement precision for reliability assessment while reducing the loss of information in terms of storage and processing requirements.
Data Source
AI summary
An apparatus comprising a memory array including a plurality of memory cells arranged in a plurality of columns and a plurality of rows is provided. The apparatus further comprises circuitry configured to perform an error detection operation on the memory array to determine a raw count of detected errors, to compare the raw count of detected errors to a threshold value to determine an over-threshold amount, to scale the over-threshold amount according to a scaling algorithm to determine a scaled error count, and to store the scaled error count in a user-accessible storage location.