Memory Data Writing Across Word Lines to Reduce Write Interference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data writing methods for rewritable non-volatile memory modules consume excessive computational resources due to the generation of parity correction codes and are prone to write interference and data loss.
Innovation Solution
A novel data writing sequence that writes multiple physical units in a different pattern across word lines, reducing resource consumption and avoiding write interference and data loss by selecting blank units strategically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If data is written sequentially to multiple physical units in the same word line according to string set order, then the writing process follows a simple pattern, but excessive computational resources are consumed in generating parity correction codes and write interference occurs
Solution Approach 1:
The patent segments the writing process by dividing physical units into different groups based on their word line and string set identifiers. Instead of writing sequentially to all physical units in the same word line, the method selects specific physical units from different word lines or different string sets within the same word line, thereby segmenting the computational workload for parity correction code generation and reducing write interference between adjacent physical units.
2Ease of operation
If data is written sequentially to all physical units in a word line before moving to the next word line, then the writing sequence is straightforward, but write interference and data loss occur
Solution Approach 1:
The patent inverts the conventional writing sequence by selecting physical units from different word lines or different string sets in an interleaved manner, rather than completing all writes in one word line before moving to the next. This inversion of the writing pattern reduces consecutive writes to adjacent physical units, thereby minimizing write interference and preventing data loss while maintaining operational simplicity through systematic selection criteria.
3Device complexity
If a conventional writing sequence is used, then the process is easy to implement, but system resources are excessively consumed during parity correction code generation
Solution Approach 1:
The patent changes the parameters of the writing sequence by introducing a selection mechanism that considers both word line identifiers and string set identifiers. Instead of using a fixed sequential pattern, the method dynamically selects physical units based on these parameters, thereby reducing the number of physical units that require simultaneous parity correction code generation. This parameter-based selection improves work efficiency while maintaining manageable implementation complexity through systematic rules.
Data Source
AI summary
A method of writing data, including: obtaining a write instruction; writing a first data of write data to a target physical unit in a target word line of a plurality of word lines in accordance with the write instruction; after writing the first data, when the target word line has one or more empty physical units arranged after the target physical unit, selecting one or more further target physical unit respectively located in one or more further target word lines; and writing second data connected after the first data in the write data to the selected further target physical units, wherein the one or more further target word lines are arranged after the target word line.