Targeted Refresh Command for Row Hammer Data Corruption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-density memory devices experience intermittent failures due to 'row hammer' issues, where repeated access to a single row causes data corruption in adjacent rows, and existing solutions like limiting access or reducing refresh time lead to performance impacts.
Innovation Solution
A targeted refresh command is sent by a memory controller to address row hammer events, identifying and refreshing victim rows to prevent data corruption without restricting access to specific memory rows, using additional pins or modifying existing commands to specify address information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of accesses per row per refresh cycle is limited, then data corruption in adjacent rows is prevented, but system performance deteriorates
Solution Approach 1:
The system performs preliminary monitoring of row access counts and proactively triggers targeted refresh commands before data corruption occurs. By detecting when a row approaches its access threshold and preemptively refreshing adjacent rows, the system prevents corruption without needing to limit overall access frequency, thus maintaining performance while ensuring data integrity.
Solution Approach 2:
Instead of applying a uniform limit to all row accesses, the system applies differentiated control only to specific rows that exhibit hammering behavior. By identifying and targeting only the affected rows and their adjacent victims for refresh operations, rather than restricting all rows equally, the system maintains high performance for normal access patterns while protecting against corruption in vulnerable areas.
2Reliability
If the refresh time is decreased, then data corruption is reduced, but performance impact increases due to additional refresh overhead
Solution Approach 1:
The system performs targeted refresh operations only on specific rows identified as victims of row hammering, rather than refreshing all rows uniformly. This localized approach reduces the total refresh overhead compared to a system-wide refresh, thereby maintaining performance while still effectively preventing data corruption in the affected areas.
Solution Approach 2:
The memory controller autonomously monitors row access patterns and automatically triggers targeted refresh commands without requiring external intervention or manual configuration. This self-service capability allows the system to dynamically respond to row hammering conditions, adjusting refresh behavior based on actual usage patterns to optimize both reliability and performance.
3Adaptability or versatility
If additional pins or command modifications are used to specify address information, then targeted refresh capability is enabled, but device complexity increases
Solution Approach 1:
The system reuses existing command structures and address lines for the targeted refresh function, making existing components serve multiple purposes. By interpreting existing commands in different contexts or utilizing already-present address bits to identify victim rows, the system achieves targeted refresh capability without requiring entirely new hardware components, thus limiting the increase in device complexity.
Data Source
AI summary
A memory controller issues a targeted refresh command. A specific row of a memory device can be the target of repeated accesses. When the row is accessed repeatedly within a time threshold (also referred to as “hammered” or a “row hammer event”), physically adjacent row (a “victim” row) may experience data corruption. The memory controller receives an indication of a row hammer event, identifies the row associated with the row hammer event, and sends one or more commands to the memory device to cause the memory device to perform a targeted refresh that will refresh the victim row.