Memory Buffer Filters Refresh Commands for Power Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory systems consume excessive power due to frequent refresh operations, which are not always necessary for maintaining data retention, especially in memory components with 'weak' cells, leading to inefficiencies in power usage and potential reliability issues.
Innovation Solution
A memory module with a command-address buffer chip that filters out a proportion of refresh commands and directs others to specific rows with 'weak' cells, allowing for reduced refresh operations while maintaining data retention fidelity, and internally generates self-refresh events at a temperature-compensated rate to optimize power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frequent refresh operations are performed to maintain data retention, then data reliability is improved, but power consumption increases
Solution Approach 1:
The patent applies local quality by differentiating refresh operations based on row characteristics. Instead of uniformly refreshing all rows, the system identifies rows with weaker cells and applies more frequent refreshes to those specific locations while reducing or skipping refreshes for rows with stronger cells. This localized approach maintains data retention reliability where needed while significantly reducing overall power consumption.
Solution Approach 2:
The patent implements partial action by performing refresh operations only on a subset of rows rather than all rows. The system calculates a refresh threshold based on row characteristics and only initiates refresh operations for rows that exceed this threshold (i.e., rows with weaker cells). This partial refresh strategy reduces the total number of refresh operations and associated power consumption while maintaining adequate data retention for critical rows.
2Reliability
If all auto-refresh commands are executed, then data retention is maintained, but power consumption is excessive
Solution Approach 1:
The patent extracts unnecessary refresh operations from the complete set of auto-refresh commands. By analyzing row characteristics and calculating refresh thresholds, the system identifies and removes (suppresses) refresh commands for rows that do not require them. Only essential refresh commands for rows with weaker cells are executed, thereby reducing overall power consumption while maintaining data retention for critical data.
Solution Approach 2:
The patent changes the parameter of refresh command execution based on row characteristics. Instead of a fixed refresh rate for all rows, the system dynamically adjusts the refresh behavior parameter for each row based on its strength characteristics. Rows with weaker cells receive more frequent refreshes, while rows with stronger cells have their refresh commands suppressed or delayed, optimizing the balance between data retention and power consumption.
3Ease of manufacture
If uniform refresh rate is used for all rows, then implementation is simple, but power consumption is inefficient
Solution Approach 1:
The patent introduces dynamics to the refresh control mechanism by making the refresh rate adaptive rather than static. The system continuously monitors or pre-characterizes row strengths and dynamically adjusts refresh behavior accordingly. This dynamic approach allows the system to optimize power consumption by matching refresh intensity to actual row requirements, while the controller manages the complexity of coordinating different refresh rates across multiple rows.
Data Source
AI summary
N out of every M number of refresh commands are ignored (filtered out) by a buffer chip on a memory module. N and M are programmable. The buffer receives refresh commands (e.g., auto-refresh commands) from the command-address channel, but does not issue a proportion of these commands to the DRAMs on the module. This reduces the power consumed by refresh operations. The buffer may replace some auto-refresh (REF) commands with activate (ACT) and precharge (PRE) commands directed to specific rows. These rows may have known ‘weak’ cells that require refreshing more often than a majority of the other rows on the module (or component). By ignoring some auto-refresh commands, and directing some others to specific rows that have ‘weak’ cells, the power consumed by refresh operations can be reduced.