Probabilistic Refresh Management for DRAM Row Hammer Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic random access memory (DRAM) devices experience data corruption due to 'row hammer' effects, where repeated activations and refreshes of a row (aggressor row) can cause neighboring rows (victim rows) to lose stored values, leading to uncorrectable and undetectable errors across multiple DRAM devices.
Innovation Solution
A memory system that issues mitigation operation commands to refresh rows within a specified vicinity of a suspected aggressor row, with a controller periodically or probabilistically adjusting the number of rows to be refreshed to mitigate data corruption, using a randomized selection technique to reconfigure the number of victim rows refreshed in response to mitigation operation commands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of victim rows refreshed is increased to mitigate row hammer effects, then data reliability is improved, but the risk of causing row hammer corruption to neighboring rows increases
Solution Approach 1:
The patent implements dynamic adjustment of the refresh scope by probabilistically changing the number of victim rows refreshed in response to mitigation operation commands. The memory device transitions from a static refresh configuration to a dynamic one where the refresh radius adapts based on operational conditions, thereby optimizing the balance between protecting against row hammer and avoiding corruption of distant rows
Solution Approach 2:
The patent changes the parameter of refresh scope (number of victim rows) probabilistically in response to mitigation commands. By adjusting this parameter dynamically rather than maintaining a fixed value, the system optimizes protection effectiveness while minimizing the risk of inducing row hammer effects on rows beyond the intended protection zone
2Reliability
If mitigation operation commands are issued frequently to refresh victim rows, then data corruption is reduced, but performance overhead increases
Solution Approach 1:
The patent implements periodic issuance of mitigation operation commands rather than continuous refreshing. By issuing commands at periodic intervals and probabilistically determining the refresh scope for each command, the system maintains data protection while reducing the overall refresh frequency and associated performance overhead compared to continuous or overly frequent refreshing
Solution Approach 2:
The patent applies partial action by refreshing only a probabilistically determined subset of potentially affected victim rows rather than all possible rows. This selective approach provides sufficient protection against row hammer while minimizing the performance impact compared to refreshing all rows in the memory array
Data Source
AI summary
A controller periodically (based on, for example, clock time or number of intervening activate commands) issues mitigation operation commands. For each (or some) mitigation operation command issued, the controller probabilistically determines whether to increase or decrease the configured number of rows in the vicinity of the suspected aggressor row that are to be refreshed by the next mitigation operation command it issues.