Stuck-Cell Memory Page Reuse via Dataset Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Memory systems face a reduction in storage capacity over time due to stuck cells being identified as 'bad' pages and discarded, leading to a shrinkage in usable memory, as conventional methods do not effectively utilize or manage stuck cells without adding them to a 'bad page' list.
Innovation Solution
A method is introduced to identify and store data in stuck-cell memory pages by matching datasets with stuck cells, allowing these pages to remain active without manipulation, and maintaining a separate 'stuck page' list to minimize the size of the 'bad page' list, thereby extending the usable life of memory arrays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If stuck cells are identified and added to the bad page list, then data integrity is maintained, but storage capacity shrinks over time
Solution Approach 1:
The patent segments the bad page list into two separate lists: a stuck page list containing pages with identified stuck cells (including stuck cell identifier and value), and a traditional bad page list for pages with other types of faults. This segmentation allows stuck pages to be reused by matching datasets to their stuck cell characteristics, thereby maintaining data integrity while preserving storage capacity.
Solution Approach 2:
The patent changes the parameter management approach by storing not just page identifiers but also specific stuck cell identifiers and stuck cell values in the stuck page list. This parameter enrichment enables the system to match datasets with corresponding stuck cell patterns, allowing stuck pages to remain in service and maintain storage capacity while ensuring data integrity through parameter-based matching.
2Reliability
If stuck cell pages are discarded to maintain reliability, then data corruption is prevented, but memory array utilization decreases
Solution Approach 1:
The patent converts the harmful effect of stuck cells into a beneficial matching mechanism. Instead of discarding pages with stuck cells, the system identifies and stores the stuck cell characteristics (identifier and value), then uses these characteristics to match and store corresponding datasets that are compatible with the stuck cell pattern. This transforms a reliability issue into a data placement opportunity, maintaining both data corruption prevention and high memory utilization.
Solution Approach 2:
The patent introduces new parameters (stuck cell identifier and stuck cell value) to characterize stuck pages, enabling a parameter-based matching approach. By storing datasets with corresponding parameter information and matching them to stuck pages with compatible parameters, the system maintains reliability through parameter verification while maximizing memory array utilization by keeping stuck pages in service.
3Reliability
If conventional bad page management is used, then faulty pages are isolated, but processing complexity increases due to list management
Solution Approach 1:
The patent segments the fault management system into distinct components: a stuck page list with specific stuck cell information, and a traditional bad page list for other faults. This segmentation allows targeted handling of different fault types, reducing overall management complexity by addressing each fault category with appropriate procedures rather than uniform bad page management.
Solution Approach 2:
The patent performs preliminary action by identifying and characterizing stuck cells early in the process, storing their identifiers and values in the stuck page list before data placement occurs. This preliminary characterization enables efficient matching and reduces subsequent processing complexity, as the system already has the necessary information to make intelligent data placement decisions without complex runtime analysis.
Data Source
AI summary
Disclosed is a method of storing data in a stuck-cell memory page of a memory array, the stuck-cell memory page having a cell at a stuck-cell-identifier cell of the memory page stuck at a stuck-cell-value, the method comprising: identifying a match between a first dataset and the stuck-cell memory page; and storing the dataset in the stuck-cell memory page; wherein the identifying a match comprises identifying that the at least one bit of the dataset which corresponds to the stuck-cell-identifier of the memory cell has a value which is equal to the stuck-cell value. A memory system configured for the above method is further disclosed