Memory Row Migration for Data Integrity in Die Cache Architectures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Die architectures face challenges in maintaining data integrity due to data errors caused by bit flips, radiation-induced errors, and noise-induced corruption, which are often addressed inefficiently with error correction mechanisms that consume valuable die real estate and resources.
Innovation Solution
A processor-based system that detects decoding errors, tracks error counts and access history for memory rows, and migrates data from high-error rows to less recently accessed rows, disabling the high-error rows temporarily to improve data integrity and reduce errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If error correction mechanisms such as ECC are employed to detect data errors, then data integrity is improved, but die real estate and system resources are consumed
Solution Approach 1:
The memory system performs self-diagnosis by monitoring its own error rates and automatically migrating data from high-error rows to low-error rows without external intervention. The processor identifies problematic memory rows through error counting and autonomously relocates data to maintain integrity while minimizing resource consumption.
Solution Approach 2:
The system dynamically changes the operational parameters of memory rows by tracking error rates and selectively disabling or migrating data from rows exceeding error thresholds. This parameter-based management allows the system to adapt to changing memory conditions and maintain reliability without permanent structural modifications.
2Reliability
If error correction mechanisms are used to detect data errors, then data integrity is improved, but device complexity increases
Solution Approach 1:
The memory system performs self-diagnosis by monitoring its own error rates and automatically migrating data from high-error rows to low-error rows without external intervention. The processor identifies problematic memory rows through error counting and autonomously relocates data to maintain integrity while minimizing resource consumption.
Solution Approach 2:
The system implements continuous feedback monitoring of error rates in each memory row, using this information to dynamically adjust data placement decisions. The error counting mechanism provides real-time feedback that drives the migration process, allowing the system to respond adaptively to changing memory conditions.
3Quantity of substance
If memory rows with high error rates are continuously used, then memory capacity is maintained, but data corruption increases
Solution Approach 1:
The system dynamically changes the operational parameters of memory rows by tracking error rates and selectively disabling or migrating data from rows exceeding error thresholds. This parameter-based management allows the system to adapt to changing memory conditions and maintain reliability without permanent structural modifications.
Solution Approach 2:
The memory system transitions from a static to a dynamic management approach, where memory row status and data placement are continuously adjusted based on real-time error rate monitoring. The system can migrate data between rows and adapt its configuration to maintain optimal reliability while preserving total memory capacity.
Data Source
AI summary
Methods and apparatuses directed to improving performance and data integrity within die architectures. In some examples, a die package includes a memory device, and a processor coupled to the memory device. The memory device may serve as a cache for another memory device. The processor receives a signal indicating that a number of errors have been detected. In response to the signal, the processor reads an error count corresponding to each of multiple memory rows of the memory device. Further, the processor determines a first memory row of the memory rows based on the error counts. The processor also determines a second memory row of the memory rows based on access data characterizing memory accesses of the plurality of rows. The processor further writes data stored at the first memory row to the second memory row of the memory device, and disables the first memory row.