Memory Error Correction Using Data Flipping for UCE Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory error correction methods fail to accurately position and correct uncorrected errors (UCE) occurring in multiple memory dice, leading to instability and inefficiency.
Innovation Solution
A method involving a register controller that performs data write and read operations to identify failure locations within a memory, using data comparison and flipping techniques to pinpoint and correct UCEs, enhancing accuracy and efficiency of error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional error correction methods are used, then corrected errors (CE) can be handled, but uncorrected errors (UCE) occurring in multiple memory dice cannot be accurately positioned or corrected
Solution Approach 1:
The patent divides the memory system into multiple memory dice and further segments the error detection process. When a UCE is detected, the system segments the search for failure location by testing each memory die individually through data write and read operations, allowing precise identification of which specific die contains the failure.
Solution Approach 2:
The patent uses data flipping techniques where test data is written in different states (original and flipped) to memory locations. By comparing read-back data with expected values, the system identifies failure locations through these state changes, analogous to using color changes for detection.
2Measurement precision
If data write and read operations are performed for each memory location to position UCE, then accurate failure location identification is achieved, but system complexity and operation time increase
Solution Approach 1:
The patent reduces system complexity by segmenting the error correction process into standardized modules: error detection, failure location identification through data write/read operations, and error correction execution. This modular approach makes the complex process more manageable and implementable.
Solution Approach 2:
The system performs self-diagnosis and self-correction by automatically detecting UCEs, identifying failure locations through systematic data operations, and correcting errors without external intervention, reducing the need for complex external control mechanisms.
3Reliability
If comprehensive data write and read operations are performed across all memory dice, then all failure locations are identified, but time consumption and operational overhead increase
Solution Approach 1:
The patent implements preliminary error detection before full error correction. By first detecting the presence of UCE and identifying affected memory dice through efficient data operations, the system prepares for targeted correction, avoiding unnecessary comprehensive operations across all memory locations.
Solution Approach 2:
The patent performs data write and read operations selectively on memory dice that are suspected of containing UCEs, rather than exhaustively testing every single memory location. This partial action approach achieves sufficient reliability while reducing time consumption compared to complete exhaustive search.
Data Source
AI summary
Example error correction methods and apparatus are described. In one example method, a register controller detects an error existing in a memory, and after detecting an uncorrected error (UCE), obtains a memory address in which the UCE occurs. The register controller reads raw data from a location indicated by the memory address, stores preset first data in the location indicated by the memory address, and reads second data from the location after storing the first data in the location. The register controller compares the first data with the second data to determine a first failure location in the location, determines raw data stored in the first failure location from the raw data in the location, and performs error correction on the raw data stored in the first failure location.