Radial Search Soft Error Detection in Memory Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory systems face challenges in detecting and correcting soft errors caused by charged particles, as these errors often accumulate in nearby memory locations, leading to multi-bit errors that are not correctable by existing error correction codes, necessitating frequent full memory scrubbing operations.
Innovation Solution
A soft error detection scheme that uses a radial search region to iteratively find and correct bit errors, adjusting the search radius based on the distance of the furthest bit error, allowing for timely correction of errors without the need for frequent full memory scrubbing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full memory scrubbing operations are performed periodically to correct soft errors, then soft errors can be corrected, but the frequency must be increased to prevent multi-bit errors from accumulating
Solution Approach 1:
The patent segments the memory array into multiple regions and uses targeted scrubbing operations on specific regions rather than performing full memory scrubbing. When a soft error is detected in one region, only that region and adjacent regions are scrubbed, leaving other regions unchanged. This segmentation allows error correction while reducing the overall scrubbing frequency and scope.
Solution Approach 2:
The patent applies local quality by focusing scrubbing operations only on the specific region where soft errors are detected, rather than uniformly scrubbing the entire memory array. The system identifies the affected region based on the detected error location and applies corrective measures locally, optimizing both reliability and productivity by avoiding unnecessary scrubbing operations in error-free regions.
2Reliability
If full memory scrubbing operations are performed frequently to prevent multi-bit errors, then error accumulation is prevented, but system performance and resource consumption increase
Solution Approach 1:
The patent divides the memory array into multiple regions and performs targeted scrubbing operations on specific regions rather than full memory scrubbing. This segmentation reduces the scope of operations requiring energy consumption, maintaining reliability while lowering resource usage.
Solution Approach 2:
The patent applies partial action by performing scrubbing operations only on the necessary regions where soft errors are detected, rather than uniformly scrubbing the entire memory array. This partial approach prevents over-engineering resource consumption while maintaining sufficient error correction capability.
3Reliability
If charged particle path is considered in soft error detection, then nearby bit errors can be found and corrected, but detection complexity increases
Solution Approach 1:
The patent applies local quality by focusing detection and correction operations on the specific region where soft errors are detected, rather than uniformly applying complex detection algorithms across the entire memory array. This localized approach reduces overall system complexity while maintaining the ability to detect and correct multi-bit errors in affected regions.
Data Source
AI summary
In a memory having a memory array, a method includes reading read data from the memory array, and detecting a first bit error in the read data. The method further includes checking all bitcells in a radial search region about the first bit error. The radial search region is defined by a search radius which indicates a number of concentric rings of bitcells physically surrounding the first bit error in the memory array.