Memory Controller Wear-Leveling for PCRAM Endurance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The increasing integration of memory devices leads to a read disturbance phenomenon, where frequent read operations on specific memory cells cause stress and data corruption, and the limited write endurance of phase change random access memory (PCRAM) devices results in reduced lifetime due to concentrated write operations.

Innovation Solution

A memory system with a memory controller that detects hot and cold memory blocks based on write and read operation counts, swapping data between these blocks to evenly distribute operations and prevent read disturbance, thereby extending the write endurance and reducing bit errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If the degree of integration of memory devices is increased, then storage capacity is improved, but read disturbance phenomenon occurs due to increased coupling effect between adjacent memory cells

Engineering Contradiction:
Improvestorage capacityVSAvoiddata integrity
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The memory device is divided into multiple memory blocks, and the controller selectively activates only the memory block containing the target address for read operations. This segmentation isolates adjacent memory cells, preventing coupling effects from affecting non-selected blocks while maintaining high integration density.

Inventive Principle:
Principle #1Segmentation

2Productivity

If write operations are concentrated on a specific memory cell region, then write speed is improved, but write endurance is reduced due to limited program cycles

Engineering Contradiction:
Improvewrite speedVSAvoidwrite endurance
Core Design Contradiction:
ProductivityVSDuration of action of stationary object

Solution Approach 1:

The controller dynamically tracks the program cycle counts of different memory blocks and adaptively selects target blocks for write operations. By continuously monitoring and adjusting the distribution of write operations based on current wear levels, the system maintains high write performance while preventing any single block from exhausting its program cycle limit.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the selection criteria for write target blocks based on program cycle history. Instead of using fixed addressing, the controller modifies block selection parameters dynamically, choosing blocks with lower program cycle counts as targets, thereby distributing wear evenly across all blocks over time.

Inventive Principle:
Principle #35Parameter changes

3Speed

If read operations are repeatedly performed on a specific memory cell, then data access speed is improved, but read disturbance causes data corruption in adjacent cells

Engineering Contradiction:
Improvedata access speedVSAvoiddata accuracy
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

By dividing the memory into isolated blocks and activating only the relevant block during read operations, the system maintains fast access speeds while preventing read disturbance from propagating to adjacent blocks. The segmentation creates electrical isolation that eliminates coupling effects during frequent read access.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS10818365B2Memory system and operating method thereof
Publication Date: 2020.10.27 MIMIRIP LLC
  • US10818365B2 patent drawing
  • US10818365B2 patent drawing
  • US10818365B2 patent drawing

AI summary

A memory system includes a memory device including a plurality of memory blocks, a first detection block suitable for detecting a hot memory block based on a number of times that a write operation is performed among the memory blocks during the write operation, a second detection block suitable for detecting first memory blocks based on the number of times that the write operation is performed among the memory blocks and detecting a cold memory block based on addresses of the first memory blocks, when the hot memory block is detected, and a wear-leveling block suitable for swapping data of the hot memory block for data of the cold memory block.