Nonvolatile Memory Cell Bloom Filters for Read Parameter Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Nonvolatile memory cells experience resistance drift over time, leading to unacceptably high error rates due to the need for unchanged read parameters, which are suboptimal at different times post-write operations, and existing timestamp-based solutions are inefficient and resource-intensive.
Innovation Solution
Implementing a plurality of Bloom filters to track write addresses, applying hash functions to each write command, and using the Bloom filters to determine elapsed time since writing, thereby selecting appropriate read parameters based on elapsed time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ECC is used to correct errors in nonvolatile memory, then data reliability is improved, but resource consumption and time required for error correction increase significantly
Solution Approach 1:
The system performs preliminary tracking of write operations using Bloom filters to identify recently written addresses before read operations occur. This allows the system to proactively select appropriate read parameters based on elapsed time since writing, preventing errors before they occur rather than correcting them after ECC processing.
Solution Approach 2:
The system uses Bloom filters to track and feedback about write operations to subsequent read operations. By maintaining information about which addresses were recently written and when, the system can adjust read parameters dynamically based on this feedback, optimizing read accuracy without requiring resource-intensive ECC processing.
2Reliability
If timestamp-based solutions are used to track write operations, then read parameter selection is improved, but device complexity and resource requirements increase
Solution Approach 1:
The invention extracts only the essential information needed for read parameter selection - whether an address was recently written - without storing complete timestamps for all addresses. By using Bloom filters to track only recently written addresses, the system removes unnecessary timestamp storage complexity while maintaining sufficient information for accurate read parameter selection.
Solution Approach 2:
Instead of storing complete timestamp data structures, the system creates a simplified copy using Bloom filters that only track the presence of recently written addresses. This copying approach maintains the necessary tracking functionality while dramatically reducing storage and processing requirements.
3Ease of operation
If read parameters are kept unchanged, then memory operation simplicity is maintained, but error rates increase due to resistance drift over time
Solution Approach 1:
The system dynamically adjusts read parameters based on elapsed time since the last write operation to each address. By making read parameters variable rather than fixed, the system adapts to resistance drift over time while maintaining simple operation through automated parameter selection based on Bloom filter tracking.
Data Source
AI summary
An apparatus includes one or more control circuit configured to connect to a nonvolatile memory cell structure. The one or more control circuit is configured to receive a plurality of write commands specifying write addresses in the nonvolatile memory cell structure. The one or more control circuit is configured to apply a plurality of Bloom filters to the write addresses, each Bloom filter corresponding to a respective period. Each Bloom filter has a plurality of hash functions to be applied to write addresses received in the respective period of the Bloom filter to generate hashes to populate the Bloom filter.