Memory Read Offset Prediction for Low-Latency QLC Reads

Solution Overview

Problem

Existing memory systems face challenges in maintaining high read performance while ensuring data integrity, particularly in QLC technology, due to the lower Read Window Budget (RWB) and the need for recalibrating read offsets with each operation, which increases latency and resource consumption.

Innovation Solution

A memory controller that selectively performs coarse and fine read level offset computations and predicts read offsets, storing them in a table for reuse, reducing the need for recalibration during each read operation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If read offsets are recalibrated with each read operation, then data integrity is maintained, but latency and resource consumption increase

Engineering Contradiction:

Improvedata integrityVSAvoidread latency

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary calibration of read offsets during initialization or when storage conditions change significantly. These pre-computed offsets are then stored and reused for subsequent read operations, eliminating the need for recalibration on every read and thereby reducing latency while maintaining data integrity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system monitors read error rates and uses this feedback to determine when recalibration is necessary. By maintaining a threshold for acceptable error rates, the system can extend the interval between calibrations, reducing the frequency of recalibration operations and thus lowering overall latency while ensuring data integrity remains within acceptable parameters.

Inventive Principle:

Principle #23Feedback

2Reliability

If read offsets are recalibrated with each read operation, then data integrity is maintained, but resource consumption increases

Engineering Contradiction:

Improvedata integrityVSAvoidresource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary calibration of read offsets during initialization or when storage conditions change significantly. These pre-computed offsets are then stored and reused for subsequent read operations, eliminating the need for recalibration on every read and thereby reducing latency while maintaining data integrity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system monitors read error rates and uses this feedback to determine when recalibration is necessary. By maintaining a threshold for acceptable error rates, the system can extend the interval between calibrations, reducing the frequency of recalibration operations and thus lowering overall latency while ensuring data integrity remains within acceptable parameters.

Inventive Principle:

Principle #23Feedback

3Measurement precision

If coarse and fine read level offset computations are performed every time, then read accuracy is maintained, but processing time increases

Engineering Contradiction:

Improveread accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system performs preliminary computation of both coarse and fine read level offsets during initialization or calibration phases. These pre-computed offset values are stored in a table structure and retrieved during subsequent read operations, eliminating the need to perform the full computational sequence on every read and thereby improving processing speed while maintaining read accuracy through the use of pre-calculated precise offset values.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The read offset computation process is segmented into two distinct phases: a calibration phase where both coarse and fine offsets are computed and stored, and an operation phase where only the fine offsets are retrieved and applied from stored tables. This segmentation allows the computationally intensive coarse offset calculation to be performed once rather than repeatedly, significantly improving processing speed while preserving read accuracy through the retained fine offset precision.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20260018216A1Memory reads using predicitve read offsets

Publication Date: 2026.01.15 MICRON TECHNOLOGY INC

AI summary

The present disclosure configures a memory sub-system controller to perform reads in a memory sub-system using predictive read offsets. The controller receives a request to read a word line group (WLG) of a set of memory components. The controller stores a table that associates a first read level offset with a first region of the WLG and a second read level offset with a second region of the WLG. The controller reads a first set of data stored in the first region of the WLG using the first read level offset stored in the table and reads a second set of data stored in the second region of the WLG using the second read level offset stored in the table.