Memory Controller Arbiter for Same-Bank Refresh Latency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Large DRAM chips face significant latency issues during refresh operations due to the need to periodically refresh all memory cells, which increases the time required to access and write data, especially in systems with several Gigabits of storage capacity, leading to increased latency and reduced performance.

Innovation Solution

The implementation of a memory controller with an arbiter that elevates the priority of memory access requests generating row activate commands during same-bank refresh operations, allowing concurrent refresh of one bank while other banks remain accessible for read and write operations, utilizing the same-bank refresh command to reduce overhead and improve bus efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional all-bank or per-bank refresh commands are used, then all memory cells are refreshed to prevent charge leakage, but the entire memory chip becomes inaccessible during the refresh interval, causing high latency

Engineering Contradiction:
Improvememory cell charge retentionVSAvoidrefresh latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The memory chip is divided into multiple independent banks, and the refresh operation is segmented to refresh only one bank at a time using same-bank refresh commands. This allows other banks to remain accessible during the refresh operation, reducing the overall refresh latency while maintaining charge retention reliability.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If larger DRAM chip sizes are used to increase storage capacity, then more memory cells are available, but the time required to refresh all memory cells increases, further increasing latency

Engineering Contradiction:
Improvestorage capacityVSAvoidrefresh time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The large memory chip is organized into multiple banks that can be refreshed independently. By using same-bank refresh commands, the system can refresh one bank while simultaneously accessing other banks, thereby maintaining high storage capacity while reducing the effective refresh time experienced by the system.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system maintains continuous useful action by allowing read and write operations to proceed in banks that are not currently being refreshed. This overlapping of refresh operations with normal memory operations ensures that the memory system remains productive throughout the refresh interval.

Inventive Principle:
Principle #20Continuity of useful action

3Reliability

If refresh operations are performed frequently to maintain charge, then data integrity is preserved, but memory access performance deteriorates due to repeated access blocking

Engineering Contradiction:
Improvedata integrityVSAvoidmemory access performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

By segmenting the memory into multiple banks and using same-bank refresh commands, the system can perform refresh operations on one bank while maintaining full access to other banks. This segmentation allows frequent refresh operations to maintain data integrity without proportionally impacting overall memory access performance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically manages bank selection for refresh operations based on current access patterns. The arbiter dynamically prioritizes memory access requests that generate row activate commands during same-bank refresh operations, adapting the refresh schedule to maintain both data integrity and performance.

Inventive Principle:
Principle #15Dynamics

4Reliability

If the arbiter prioritizes refresh commands over memory access requests, then charge leakage is prevented, but pending read and write commands experience increased latency

Engineering Contradiction:
Improvecharge retentionVSAvoidcommand latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The arbiter uses segmentation by bank to allow memory access requests to be serviced in banks that are not currently being refreshed. By prioritizing access requests that can proceed in available banks while scheduling refresh operations for other banks, the system prevents charge leakage without unnecessarily delaying pending commands.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS12118247B2Performance of bank refresh
Publication Date: 2024.10.15 ADVANCED MICRO DEVICES INC
  • US12118247B2 patent drawing
  • US12118247B2 patent drawing
  • US12118247B2 patent drawing

AI summary

A memory controller includes an arbiter. The arbiter is configured to elevate a priority of memory access requests that generate row activate commands in response to receiving a same-bank refresh request, and to send a same-bank refresh command in response to receiving the same-bank refresh request.