DRAM Refresh Control for Row Hammer Victim Rows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Volatile memory devices like DRAM experience data loss due to charge leakage and row hammering, where intense access to certain rows affects adjacent rows, necessitating improved refresh operations.
Innovation Solution
A memory device with a refresh controller that generates control signals to identify aggressor and victim row addresses, determining whether to refresh or skip refresh operations based on row address differences, thereby reducing power consumption and enhancing defense performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If refresh operations are performed on all rows periodically, then data integrity is maintained, but power consumption increases and row hammering vulnerability persists
Solution Approach 1:
The patent applies local quality by performing refresh operations selectively on specific victim rows adjacent to aggressor rows rather than uniformly refreshing all rows. The refresh controller identifies aggressor rows through intensive access patterns and targets only the adjacent victim rows for refresh, making the refresh operation localized to where it is actually needed. This resolves the contradiction by maintaining data integrity in vulnerable areas while avoiding unnecessary power consumption from refreshing already-safe rows.
Solution Approach 2:
The patent implements preliminary action by proactively identifying aggressor rows through monitoring access patterns and pre-refreshing adjacent victim rows before charge leakage can cause data loss. The refresh controller detects rows with intensive access (aggressor rows) and performs refresh operations on neighboring rows in advance, preventing potential data corruption rather than reacting after damage occurs. This approach maintains reliability while optimizing power usage by targeting only at-risk rows.
2Reliability
If refresh operations are performed frequently on all rows, then data integrity is maintained, but defense performance against row hammering is insufficient
Solution Approach 1:
The patent applies local quality by directing refresh operations specifically to victim rows adjacent to identified aggressor rows rather than uniformly refreshing all rows. The refresh controller monitors access patterns to identify aggressor rows and selectively refreshes only the neighboring victim rows that are vulnerable to row hammering effects. This localized approach enhances defense performance by concentrating refresh resources on actually vulnerable rows while avoiding waste on already-safe rows.
Solution Approach 2:
The patent implements feedback by continuously monitoring row access patterns to identify aggressor rows and using this information to dynamically adjust refresh operations. The refresh controller receives feedback about which rows are being intensively accessed and responds by targeting adjacent rows for refresh. This closed-loop approach enhances defense performance against row hammering by adapting refresh operations based on real-time access pattern analysis.
3Use of energy by moving object
If selective refresh based on aggressor row identification is implemented, then power consumption is reduced, but device complexity increases
Solution Approach 1:
The patent applies universality by designing the refresh controller to perform multiple functions: it monitors row access patterns to identify aggressor rows, determines adjacent victim rows, and executes selective refresh operations. This multi-functional approach consolidates what could be separate complex subsystems into a single integrated controller, reducing overall device complexity while enabling power-efficient selective refresh operations.
Solution Approach 2:
The patent implements self-service by enabling the refresh controller to autonomously identify aggressor rows through access pattern monitoring and automatically determine which victim rows require refresh without external intervention. The system serves itself by using its own access pattern data to make refresh decisions, eliminating the need for additional complex control logic or external coordination while achieving power-efficient selective refresh.
Data Source
AI summary
A memory device may include a refresh controller configured to generate a first control signal and a second control signal based on an activate command and a row address that corresponds to the activate command, an aggressor row determiner configured to determine the row address as an aggressor row address based on the first control signal, and a victim row determiner configured to determine a victim row address based on the aggressor row address, and determine whether to output the victim row address as a refresh address based on the second control signal. The memory device may be configured to perform a refresh on a row corresponding to the refresh address.