Memory Row Classification Circuit for Row Hammering Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As memory integration increases, the coupling effect between adjacent word lines leads to data damage due to row hammering, where frequently activated word lines disturb adjacent memory cells, and in cryogenic memory, the reduced need for refresh operations exacerbates data loss from this phenomenon.
Innovation Solution
A memory system with a target row classification circuit and signal generation circuit that identifies susceptible rows and performs precharge and activation operations to mitigate data loss, including additional refresh operations when necessary, even in cryogenic environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the space between word lines is reduced to increase integration, then the degree of integration is improved, but the coupling effect between adjacent word lines increases causing data damage
Solution Approach 1:
The patent applies preliminary action by counting the number of times each word line has been activated and using this information to predict which word lines are likely to cause row hammering. The system proactively identifies susceptible rows before data damage occurs and performs targeted refresh operations to prevent the harmful coupling effect from manifesting.
2Reliability
If additional refresh operations are performed to prevent row hammering, then data reliability is improved, but device complexity increases
Solution Approach 1:
The patent applies local quality by differentiating between regular refresh operations and targeted additional refresh operations. Instead of uniformly refreshing all rows, the system identifies specific word lines and rows that are susceptible to row hammering based on activation counts and coupling effects, then applies enhanced refresh operations only to those local areas, thereby improving data reliability without unnecessarily increasing overall device complexity.
Solution Approach 2:
The patent implements feedback by continuously monitoring the number of activations for each word line and using this information to dynamically adjust refresh strategies. The system uses activation count data to determine which rows require additional refresh operations, creating a closed-loop control mechanism that adapts to changing operational conditions and prevents row hammering while optimizing resource usage.
3Reliability
If refresh operations are performed frequently to prevent data loss, then data integrity is improved, but energy consumption increases
Solution Approach 1:
The patent applies partial action by performing refresh operations only on the subset of rows that are identified as susceptible to row hammering, rather than refreshing all rows uniformly. This selective approach maintains data integrity for the critical rows that are most at risk while avoiding unnecessary energy consumption from refreshing rows that do not require it, thus optimizing the balance between data integrity and energy consumption.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively reduces data loss from row hammering by proactively identifying and refreshing target rows, ensuring data integrity even in systems where regular refresh operations are infrequent, such as those operating at cryogenic temperatures.
Implementation Method 1
As the coupling effect between adjacent word lines increases as described above, there occurs a phenomenon in which data of a memory cell coupled to a word line adjacent to a frequently activated word line is damaged
Data Source
AI summary
A memory may include multiple rows each coupled to multiple memory cells; a target row classification circuit suitable for classifying, as a target row, a row, among the multiple rows, that is susceptible to data loss as a result of activity of an adjacent row; and a target row signal generation circuit suitable for sequentially activating a target row active signal for activating the target row and a target row precharge signal for precharging the target row in response to a precharge command.