Memory Subsystem Controller Read Throughput Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D cross-point memory devices experience significant read voltage threshold (VT) shifts during high read throughput conditions, leading to read failures and data loss due to power supply limitations and high concurrency, without a mechanism to manage the VT drop.
Innovation Solution
A memory sub-system controller detects read errors during high throughput and adjusts the read throughput level by limiting queue depth or command spacing, executing a read retry operation at a reduced throughput to recover data, and then returns to the original throughput level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If read throughput is increased to improve data access speed, then productivity is improved, but read voltage threshold shifts occur leading to read failures and data loss
Solution Approach 1:
The system dynamically adjusts the read throughput level based on detected read errors. When read failures occur during high throughput operations, the controller automatically lowers the throughput level to a reduced level for retry operations, then restores the original throughput level after successful recovery. This dynamic adaptation resolves the contradiction by allowing high productivity during normal operation while ensuring reliability when errors occur.
Solution Approach 2:
The invention changes the operational parameter (read throughput level) in response to detected read errors. By transitioning from a high throughput level to a reduced throughput level during error recovery, and then returning to the original level after successful retry, the system resolves the contradiction between maintaining high productivity and ensuring read operation reliability under varying conditions.
2Reliability
If queue depth is limited to reduce read throughput for data recovery, then read operation reliability is improved, but productivity decreases
Solution Approach 1:
The system dynamically adjusts queue depth based on operational conditions. During normal operation, queue depth is maintained at high levels for maximum productivity. When read errors are detected, the system temporarily limits queue depth to reduce throughput and enable successful data recovery operations, then restores the original queue depth configuration after recovery completes.
Solution Approach 2:
The invention implements periodic adjustment of queue depth in response to error detection. The system operates at high throughput with full queue depth, detects read errors, temporarily reduces queue depth for recovery operations, then returns to normal operation. This periodic adjustment between high-performance mode and recovery mode resolves the contradiction between reliability during recovery and productivity during normal operation.
3Reliability
If command spacing is increased to lower read throughput, then read operation reliability is improved, but access time increases
Solution Approach 1:
The system dynamically adjusts command spacing intervals based on operational phase. During normal high-throughput operation, command spacing is minimized for fast access. When read errors occur during recovery operations, the system increases command spacing to reduce throughput and improve read retry success probability, then returns to minimal spacing after successful recovery.
Solution Approach 2:
The invention implements periodic modification of command spacing parameters. The system operates with minimal command spacing for maximum speed, detects read errors, temporarily increases command spacing to enable successful recovery reads, then restores minimal spacing for normal operation. This periodic adjustment resolves the contradiction between access time during normal operation and read retry success during recovery.
Data Source
AI summary
An error associated with a read operation corresponding to a memory die of a memory sub-system is detected. In response to detecting the error, a first read throughput level of the memory sub-system is identified. A quantity of queues receiving operation requests is decreased, the decreased quantity of queues corresponding to a second read throughput level. A read retry operation associated with the memory die is initiated at the second read throughput level.