DRAM Refresh Management via Per-Row Frequency Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic Random Access Memory (DRAM) systems waste resources due to a fixed refresh period that is determined by the weakest memory cell, leading to excessive refreshes of many rows, as the refresh requirements vary significantly among cells, resulting in inefficient use of system resources.
Innovation Solution
Implementing a memory controller that cooperates with DRAM chip logic to manage refresh operations by storing a memory table with row addresses and their required refresh rates, allowing for targeted extra refresh cycles to accommodate weak cells and rows, thereby optimizing refresh frequency and reducing unnecessary refresh commands across all DRAM banks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed refresh period is specified for DRAM, then all rows are refreshed uniformly, but this causes many rows to be refreshed more frequently than required, wasting system resources
Solution Approach 1:
The patent segments the memory system into individual row units, each with its own refresh counter and control logic. This allows each row to be managed independently based on its specific refresh requirements, rather than forcing all rows to follow a uniform refresh schedule. The segmentation enables targeted refresh operations that match actual data retention needs of each row.
Solution Approach 2:
The patent implements dynamic refresh period adjustment for each row based on its specific retention characteristics. Instead of a fixed refresh period for all rows, the system dynamically determines optimal refresh intervals for individual rows, allowing rows with better retention to be refreshed less frequently while maintaining data integrity for rows with poorer retention.
2Reliability
If the weakest memory cell determines the refresh period for the entire DRAM, then all cells are refreshed, but this causes most cells to be refreshed more frequently than necessary
Solution Approach 1:
The patent applies local quality by assigning different refresh characteristics to different rows based on their specific needs. Each row can have its own refresh period tailored to its retention properties, rather than applying a single uniform refresh period determined by the weakest cell. This localized approach optimizes system efficiency while maintaining reliability for each individual row.
Solution Approach 2:
The patent changes the refresh period parameter individually for each row based on measured retention characteristics. By monitoring and adapting the refresh period for each row separately, the system avoids the inefficiency of using the weakest cell's parameters for all rows, thereby improving overall system productivity while maintaining data retention.
3Ease of operation
If refresh operations are performed on a per row basis, then individual row control is possible, but the row with the weakest cell determines the refresh period for all rows
Solution Approach 1:
The patent implements self-service by enabling each row to effectively determine its own refresh requirements through individual monitoring and control logic. Each row's refresh operations are autonomously managed based on its specific retention characteristics, eliminating the need for one row to dictate the refresh schedule for all other rows. This self-service approach reduces excessive refresh operations while maintaining ease of per-row control.
Data Source
AI summary
Apparatus, systems, and methods to manage memory refresh operations are described. In one embodiment, an electronic device comprises a processor and memory controller logic to determine a memory refresh frequency for a memory system and transmit refresh commands to a refresh control logic in at least one memory bank coupled to the memory controller according to the memory refresh frequency. Other embodiments are also disclosed and claimed.