Adaptive Memory Partitioning for NUTS Data Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory sub-systems inefficiently manage partitions by not distinguishing between types of data being stored, leading to increased PE cycles, wasted space, and reduced memory system life due to small partitions of non-user targeted space (NUTS) data.
Innovation Solution
Adaptive read level threshold operations are implemented by a memory sub-system controller to condition partition closure based on data type, delaying partition creation for NUTS data and associating it with a default read level until host data is stored, thereby reducing PE cycles and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If partitions are created for all data including NUTS data, then data is organized into manageable units, but the number of PE cycles increases and memory system life decreases
Solution Approach 1:
The patent applies local quality by treating NUTS data and host data differently in terms of partition management. NUTS data regions are excluded from partition creation while host data regions receive normal partition management, allowing localized optimization without affecting the entire memory system. This selective approach reduces unnecessary PE cycles on NUTS data while maintaining data organization where needed.
Solution Approach 2:
The patent segments the memory space into distinct regions based on data type - NUTS data regions and host data regions. By identifying and separating these regions, the system can apply different management strategies: no partitioning for NUTS data (reducing PE cycles) and normal partitioning for host data (maintaining productivity). This segmentation resolves the contradiction by applying partitioning only where beneficial.
2Productivity
If partitions are created for NUTS data, then data is organized into manageable units, but space is wasted due to small partition sizes
Solution Approach 1:
The patent extracts NUTS data from the normal partitioning流程, removing it from the partition management system entirely. By taking out NUTS data regions and excluding them from partition creation, the system eliminates the problem of small, inefficient partitions for non-user data while maintaining space availability for future host data allocations.
Solution Approach 2:
Instead of creating partitions for all data and then managing them, the patent inverts the approach by deliberately not creating partitions for NUTS data. This inversion eliminates the overhead and space waste associated with small partitions while still allowing the memory to be used efficiently for its intended purpose - storing non-user targeted space data.
3Device complexity
If partitions are closed based on fixed time intervals, then partition management is simplified, but efficiency is reduced due to premature closure
Solution Approach 1:
The patent introduces dynamic partition management by monitoring the type of data being written to memory regions. Instead of fixed time-based closure, partitions are dynamically adjusted based on whether host data or NUTS data is being stored. This dynamic approach allows partitions to remain open longer when NUTS data is detected, improving efficiency without significantly increasing management complexity.
Solution Approach 2:
The patent implements feedback mechanisms by monitoring data type information during the write process. This feedback allows the partition management system to make informed decisions about whether to create or close partitions based on the actual data being stored, rather than following a predetermined schedule. The feedback loop enables efficient adaptation to varying data patterns while maintaining manageable complexity through automated detection.
Data Source
AI summary
Aspects of the present disclosure configure a system component, such as a memory sub-system controller, to perform adaptive read level threshold voltage operations. The controller receives a request to program data into an individual portion of the set of memory components and determines whether the request comprises host data or non-user targeted space (NUTS) data. The controller conditionally defines a partition for the individual portion of the set of memory components based on whether the request comprises host data or the NUTS data. The controller associates the partition with a bin of a plurality of bins, each of the plurality of bins representing an individual read level offset used to access a charge distribution of data stored in the individual portion of the set of memory components.