Memory Subsystem Pseudo Media Management for Computational Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory sub-systems face performance degradation due to computationally expensive operations required for determining actual health characteristic values, such as minimum erase cycle counts, which are necessary for media management operations like static wear leveling.
Innovation Solution
The implementation of pseudo media management operations that can be selectively performed without determining actual health characteristic values, thereby reducing computational overhead and performance degradation. These pseudo operations increment health characteristic values in response to memory operations, allowing for delayed actual media management until necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If actual health characteristic values are determined for media management operations, then media management accuracy is improved, but computational overhead increases and performance degrades
Solution Approach 1:
The system performs preliminary determination of health characteristic values during normal operation and stores them for later use. When media management operations are needed, the pre-determined values are used instead of calculating actual values in real-time, thus avoiding performance degradation while maintaining management accuracy.
Solution Approach 2:
The system creates copies of health characteristic values and stores them in a data structure for efficient retrieval. Instead of repeatedly calculating actual health values during media management operations, the system uses these pre-stored copies, reducing computational overhead while preserving the accuracy needed for proper media management.
2Reliability
If health characteristic values are determined frequently, then media management accuracy is improved, but computational overhead increases
Solution Approach 1:
The system determines health characteristic values periodically rather than continuously or on every media management operation. A data structure stores these periodically determined values and makes them available for multiple media management operations, reducing the frequency of expensive calculations while maintaining sufficient reliability for media management tasks.
Solution Approach 2:
Health characteristic values are determined in advance and stored in a readily accessible data structure before media management operations occur. This preliminary determination allows multiple management operations to use the same pre-calculated values without repeating the computationally expensive determination process, thus reducing energy consumption while maintaining management reliability.
3Speed
If actual health characteristic values are determined in real-time, then responsiveness is improved, but system performance degrades
Solution Approach 1:
The system determines and stores health characteristic values in advance in a data structure, making them immediately available for media management operations. This eliminates the need for real-time calculation during management operations, maintaining fast responsiveness while preserving overall system throughput by avoiding repeated expensive computations.
Solution Approach 2:
The system creates and stores copies of health characteristic values that can be rapidly retrieved during media management operations. These pre-stored copies provide the responsiveness needed for timely media management while preventing the performance degradation that would result from calculating actual values in real-time during each operation.
Data Source
AI summary
A method includes determining a gap between a difference in a first health characteristic value and a second health characteristic value of blocks of memory cells and a health threshold associated with the blocks of memory cells, determining the gap is greater than or equal to a gap threshold from the health threshold, performing a pseudo media management operation on the blocks of memory cells, and determining an updated first health characteristic value of the blocks of memory cells.