Memory Bank Group Architecture for Concurrent Refresh and Precharge
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional bank architectures in memory devices do not scale well with increased size and speed, leading to inefficiencies in data I/O bandwidth utilization, particularly due to the need for increased core cycle times which are power and die-size intensive, and the high operational bandwidth consumption by refresh commands in volatile memory devices.
Innovation Solution
Implementing a memory subsystem that allows for refreshing and precharging of one or more identified banks across different bank groups with a single command, enabling concurrent operations across multiple bank groups, thereby reducing the number of commands required and improving bandwidth utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional bank architecture is used to increase memory size and speed, then data I/O bandwidth is improved, but core cycle time increases which leads to increased power consumption and die size
Solution Approach 1:
The memory device segments banks into multiple bank groups, allowing independent access to different bank groups. This segmentation enables parallel operations where the memory controller can access one bank group while another bank group performs refresh operations, thereby maintaining high data I/O bandwidth without increasing core cycle time and reducing power consumption.
2Productivity
If bank group architecture is used to improve data bandwidth utilization, then access timing across banks is improved, but the number of refresh commands increases consuming operational bandwidth
Solution Approach 1:
The patent merges the refresh operations of multiple banks across different bank groups into a single unified refresh command. When a refresh command is issued, all banks in all bank groups are refreshed simultaneously, reducing the number of refresh commands from potentially dozens (one per bank) to just one or a few commands, thereby freeing up operational bandwidth for data transfers.
Solution Approach 2:
The refresh command structure is designed to be universal, applying to all banks across all bank groups simultaneously. This multi-functional refresh command can target any combination of banks without requiring separate commands, simplifying the command interface and reducing the complexity of refresh management while maintaining high data bandwidth utilization.
3Quantity of substance
If more banks are added to increase memory capacity, then memory size is improved, but the number of commands required to maintain memory increases
Solution Approach 1:
By organizing banks into bank groups, the system can address and control multiple banks using fewer command structures. The bank group architecture allows the memory controller to issue commands at the bank group level or selectively to individual banks within groups, reducing the total number of commands needed to manage large numbers of banks while maintaining high memory capacity.
Solution Approach 2:
The patent combines multiple bank management operations into unified commands. Instead of requiring separate commands for each bank, the system uses merged commands that can operate on multiple banks simultaneously or on-demand, thereby scaling memory capacity without proportionally increasing command complexity.
Data Source
AI summary
Memory subsystem refresh management enables commands to access one or more identified banks across different bank groups with a single command. Instead of sending commands identifying a bank or banks in separate bank groups by each bank group individually, the command can cause the memory device to access banks in different bank groups. The command can be a refresh command. The command can be a precharge command.