Memory Read Error Handling Using Bin-Based Corrective Reads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory sub-systems face performance degradation due to slow charge loss (SCL) leading to increased read error rates and latency, particularly in memory cells with longer data retention, as conventional read error handling (REH) flows are not optimized for varying cell states.
Innovation Solution
Implementing bin-based read error handling (REH) flows with a fast corrective read (CR) operation tailored to the classification of memory cells, using different REH flows based on bin classification to mitigate temporal voltage shift (TVS) and reduce latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional read error handling flows are used for all memory cells, then read error handling is performed uniformly, but latency increases and performance degrades due to slow charge loss in cells with longer data retention
Solution Approach 1:
The patent segments memory cells into different bins based on their data retention characteristics and temporal voltage shift profiles. By dividing the memory cell population into distinct groups (bins) with similar characteristics, the system can apply tailored read error handling flows to each bin, avoiding the latency penalty of uniform conservative handling across all cells.
Solution Approach 2:
The patent implements dynamic selection of read error handling flows based on the bin classification of memory cells. The system adapts the error handling approach in real-time by selecting appropriate flows according to the specific retention characteristics of each cell bin, optimizing latency while maintaining reliability.
2Loss of time
If optimized fast corrective read operations are applied to all memory cells, then latency is reduced, but read error rates increase due to insufficient error handling for cells with longer data retention
Solution Approach 1:
The patent applies different quality levels of error handling to different memory cell bins. Cells in bins with shorter data retention receive faster, less conservative error handling, while cells in bins with longer retention receive more robust error handling. This local differentiation optimizes overall system performance by matching error handling intensity to actual cell characteristics.
Solution Approach 2:
The patent changes the parameters of read error handling operations based on bin classification. Different read error handling flows with varying parameters (such as read voltage levels, retry counts, and correction algorithms) are applied to different bins, allowing the system to optimize latency for cells that need it while maintaining reliability for cells that require more protection.
3Productivity
If bin-based read error handling flows are implemented, then read error handling is optimized for varying cell states, but device complexity increases
Solution Approach 1:
While segmentation into bins increases complexity, it enables the system to manage complexity through organization. By grouping cells with similar characteristics into bins, the system reduces the complexity of managing individual cell variations and enables more efficient error handling strategies.
Solution Approach 2:
The patent creates multiple read error handling flows that can be universally applied to different bins. Each flow is designed to handle specific characteristics of memory cell bins, and the same set of flows can be reused across different bin types, reducing the overall complexity compared to creating custom handling logic for each individual cell or bin combination.
Data Source
AI summary
A system comprises a memory device and a processing device, operatively coupled with the memory device. The processing device detects an error during a read operation on a set of target cells of the memory device. The processing device selects, based on a state information bin with which the set of target cells is associated, a read error handling (REH) flow for the set of target cells, wherein the set of target cells is associated with the state information bin based on corresponding cell state information. The processing device executes the REH flow on the set of target cells of the memory device to correct the error.