Deep Learning Read Threshold Estimation for Memory Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems face challenges in optimizing read threshold values due to varying operating conditions, leading to read errors and performance degradation, as manual adjustment is impractical and inefficient.
Innovation Solution
A deep learning-based system that classifies operating conditions contributing to read errors and determines optimal read threshold sets by performing multiple read operations, obtaining fail bit count information, and predicting the best read threshold set using deep neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual adjustment of read threshold values is performed, then read errors can be reduced, but the process becomes impractical and inefficient
Solution Approach 1:
The memory system performs self-diagnosis and self-adjustment of read threshold values through automated algorithms that analyze read errors and dynamically optimize thresholds without external manual intervention, making the system self-sufficient in maintaining optimal performance
Solution Approach 2:
The system dynamically changes read threshold parameters based on analyzed error patterns and operating conditions, automatically adjusting threshold values to optimize read operations under varying conditions without manual reconfiguration
2Reliability
If multiple read operations are performed to determine optimal read thresholds, then read performance is improved, but processing time increases
Solution Approach 1:
The system performs preliminary characterization of read error patterns and determines optimal read thresholds in advance through initial multiple read operations, storing these optimized thresholds for subsequent use to avoid repeated time-consuming measurements during normal operation
Solution Approach 2:
The system implements feedback mechanisms where read error information from multiple operations is analyzed and used to iteratively refine and update read threshold values, reducing the need for extensive repeated measurements by learning from previous results
Data Source
AI summary
A controller estimates optimal read threshold values for a memory device using deep learning. The memory device includes multiple pages coupled to select word lines in a memory region. The controller performs multiple read operations on a select type of page for each word line using multiple read threshold sets, obtains fail bit count (FBC) information associated with each read operation, and determines an optimal read threshold set for each word line based on the FBC information. When optimal read threshold sets for the select word lines are different each other, the controller predicts a best read threshold set using the optimal read threshold sets.