DDR4 Memory C/A Signal Training via Parity Error Feedback

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In memory systems, especially those using DDR3 interfaces, errors in command-address (C/A) signals with respect to clock signals lead to prolonged boot-up times and inefficiencies due to the time-consuming JEDEC standardized initialization process required to restore functionality.

Innovation Solution

A method and apparatus for training C/A signals using a C/A parity error signal, allowing for rapid retraining without initiating a full system reboot or JEDEC initialization, utilizing a training logic unit that adjusts the phase of C/A signals based on parity error feedback to maintain system functionality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the JEDEC standardized initialization process is used to resolve memory module errors, then the memory module is restored to a known functional state, but the system experiences prolonged boot-up time and performance loss

Engineering Contradiction:
Improvememory module functionalityVSAvoidboot-up time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the error resolution process into two distinct paths: a fast retraining path for C/A signal errors and a full initialization path for other errors. This segmentation allows the system to apply the appropriate resolution method based on error type, avoiding unnecessary full initializations and reducing boot-up time while maintaining reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements preliminary error detection and classification during the memory training phase. By detecting C/A parity errors early and classifying them before full initialization occurs, the system can prepare to execute the faster retraining sequence, preventing the need for time-consuming full initialization and thereby reducing boot-up time.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If the memory controller delays the phase of C/A signals during training to achieve correct data reception, then data accuracy is improved, but the training process becomes time-consuming and may lead to unknown states

Engineering Contradiction:
Improvedata reception accuracyVSAvoidtraining time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent employs feedback through C/A parity error signals during the training phase. The memory controller monitors these error signals in real-time and uses them to adjust the phase of C/A signals dynamically. This feedback mechanism enables the system to achieve correct data reception accuracy while minimizing training time by directly responding to error conditions rather than attempting all possible phase combinations.

Inventive Principle:
Principle #23Feedback

3Reliability

If the memory controller implements continuous monitoring of C/A parity error signals, then error detection capability is improved, but the system complexity increases

Engineering Contradiction:
Improveerror detection capabilityVSAvoidtraining logic unit complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary C/A parity error signal as a mediator between the memory module and the training logic unit. This intermediary signal simplifies the monitoring process by providing a direct error indication that the training logic unit can act upon, reducing the complexity of continuous monitoring while maintaining high error detection capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2586031B1Method and apparatus for training a memory signal via an error signal of a memory
Publication Date: 2018.09.12 INTEL CORP
  • EP2586031B1 patent drawingFigure 1
  • EP2586031B1 patent drawingFigure 2
  • EP2586031B1 patent drawingFigure 3

AI summary

Described herein is a method and an apparatus for training a memory signal via an error signal of a memory. The method comprises transmitting from a memory controller a command-address (C/A) signal to a memory module; determining by the memory controller an error in the memory module via an error signal from an error pin of the memory module, the error associated with the C/A signal transmitted to the memory module; and modifying by the memory controller the C/A signal in response to determining an error in the memory module, wherein the error pin is a parity error pin of the memory module, and wherein the memory module comprises a Double Data Rate 4 (DDR4) interface.