Multi-dimensional Data Randomization in Memory Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Certain memories experience data loss due to physically-proximate data bits having the same value, leading to inter-cell interference, which existing data randomization techniques partially address by altering the mix of ones and zeros.
Innovation Solution
Implementing data randomization in multiple dimensions by using a first scrambling sequence for one row of a memory array and a circularly shifted version of this sequence for adjacent rows, ensuring even randomization along rows and columns, and potentially in a third dimension for 3D arrays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If data randomization is applied using a single scrambling sequence for all rows, then the implementation is simple, but data loss occurs due to physically-proximate bits having the same value
Solution Approach 1:
The patent extends data randomization from a single-dimension approach (applying the same scrambling sequence to all rows) to a multi-dimensional approach by using circularly shifted versions of the scrambling sequence for different rows. This dimensional extension ensures that physically-proximate bits across adjacent rows have different values, reducing inter-cell interference and improving data retention while maintaining implementation simplicity through the systematic use of sequence shifting.
2Ease of operation
If the same scrambling sequence is used for all rows, then the implementation is straightforward, but inter-cell interference occurs between adjacent rows
Solution Approach 1:
The patent applies local quality by making the scrambling sequence locally unique to each row through circular shifting. Each row receives a scrambled version of the data using a scrambling sequence that is shifted by a specific amount relative to other rows, creating local differentiation that eliminates inter-cell interference while maintaining overall system simplicity.
3Productivity
If data is not randomized, then the storage process is efficient, but wear is unevenly distributed across memory cells
Solution Approach 1:
The patent applies preliminary action by randomizing data with circularly shifted scrambling sequences before writing to memory cells. This pre-processing step ensures that zeros and ones are evenly distributed across all memory cells including spares, preventing any single cell from experiencing excessive wear from storing repeated patterns, thereby extending overall memory lifespan while maintaining storage efficiency.
Data Source
AI summary
The disclosed technology provides for multi-dimensional data randomization in a memory cell array using circular shifts of an initial scrambling sequence. Data addressed to a first row of a data array is randomized using the initial scrambling sequence and data addressed to each row of the memory cell array is randomized using a scrambling sequence that is equal to a circular shift of the initial sequence.