Memory Wordline ECC Detection for Failing Array Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current error detection schemes in memory systems, particularly in cache designs, often fail to detect wordline errors effectively, leading to mis-correction and uncorrectable errors, especially when multiple bits fail simultaneously, which is unacceptable in server systems.

Innovation Solution

A controller is implemented to read multiple error correcting code (ECC) words from a wordline and detect errors by identifying a failing wordline when more than one error is detected across these ECC words, using a restricted approach to prevent false errors and mis-correction, thereby enhancing error detection and correction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional error detection schemes are used in memory systems, then the device complexity is reduced, but the reliability of error detection deteriorates leading to mis-correction and uncorrectable errors

Engineering Contradiction:
Improveerror detection reliabilityVSAvoiddetection circuitry complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the error detection process into multiple independent syndrome checks across different ECC words within the same wordline. Instead of treating each ECC word independently, the system divides the wordline into multiple segments (ECC words) and applies error detection logic to each segment separately, then combines the results to identify wordline-level failures.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a nested error detection structure where multiple levels of error checking are combined. The first level checks individual ECC words for correctable errors using traditional syndrome decoding. The second level nests within this by checking multiple ECC words from the same wordline to detect uncorrectable errors. This nested approach allows the system to detect both correctable and uncorrectable errors without requiring completely separate detection circuits.

Inventive Principle:
Principle #7Nested doll (Nesting)

2Reliability

If multiple bits fail simultaneously in a wordline, then the data corruption risk increases, but traditional error correction schemes fail to detect these gross errors

Engineering Contradiction:
Improvedata integrityVSAvoidgross error detection difficulty
Core Design Contradiction:
ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent merges the error detection results from multiple ECC words that share the same wordline. By combining the syndrome information from multiple ECC words, the system can detect gross errors that affect the wordline. When multiple ECC words from the same wordline exhibit error patterns, the merged analysis identifies this as a wordline-level failure rather than independent bit errors, enabling detection of gross errors that would otherwise be mis-corrected.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent implements feedback mechanisms where the detection of errors in one ECC word influences the interpretation of errors in subsequent ECC words from the same wordline. When a correctable error is detected in an ECC word, the system feeds this information back to the wordline error detection logic. If similar error patterns appear in other ECC words from the same wordline, the feedback mechanism triggers a wordline failure identification, preventing mis-correction of gross errors.

Inventive Principle:
Principle #23Feedback

3Reliability

If a restricted approach is used to prevent false errors and mis-correction, then the reliability improves, but the detection capability for different error types may be limited

Engineering Contradiction:
Improvefalse error preventionVSAvoiderror type detection versatility
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent applies different error detection and correction strategies to different local contexts within the memory system. For individual ECC words, traditional SEC/DED (Single Error Correction, Double Error Detection) logic is applied. For multiple ECC words sharing the same wordline, a different detection logic is applied that looks for patterns indicating wordline failures. This local quality approach allows the system to maintain high reliability for typical bit errors while also detecting gross wordline errors, without a single approach limiting versatility.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent creates a universal error detection framework that handles multiple error types through a unified wordline analysis mechanism. The same detection circuitry that identifies correctable errors in individual ECC words also serves to detect uncorrectable errors when analyzing multiple ECC words from the same wordline. This multi-functional approach allows a single detection system to handle both correctable and uncorrectable errors, as well as bitline and wordline failures, without requiring separate specialized circuits for each error type.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS8914708B2Bad wordline/array detection in memory
Publication Date: 2014.12.16 GLOBALFOUNDRIES US INC
  • US8914708B2 patent drawing
  • US8914708B2 patent drawing
  • US8914708B2 patent drawing

AI summary

A technique for error detection is provided. A controller is configured to detect errors by using error correcting code (ECC), and a cache includes independent ECC words for storing data. The controller detects the errors in the ECC words for a wordline that is read. The controller detects a first error in a first ECC word on the wordline and a second error in a second ECC word on the wordline. The controller determines that the wordline is a failing wordline based on detecting the first error in the first ECC word and the second error in the second ECC word.