Memory Controller Reliability Data Using Syndrome Weight
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory systems face challenges in executing robust error correction for read data, particularly when error bits are present, as they often rely on hard-decision decoding which may fail, necessitating a more effective method to determine reliability and correct errors accurately.
Innovation Solution
A controller and operating method that utilize storage memory to store read data and reliability data units, a decoder to execute decoding operations based on reliability data, and a processing circuit to calculate syndrome weights from read data under different biases, determining reliable data units by comparing these weights and their differences to select the most accurate read data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hard-decision decoding is used for error correction, then the decoding process is simple and fast, but error correction reliability is insufficient and may fail when error bits are present
Solution Approach 1:
The patent performs preliminary actions by calculating syndrome weights and determining reliability data units before executing the decoding operation. The controller calculates syndrome weights for multiple read data candidates, determines reliability information for each bit position in advance, and then uses this pre-computed reliability data to guide the soft-decision decoding process, improving error correction reliability without excessive complexity during the actual decoding
Solution Approach 2:
The patent introduces reliability data units as an intermediary between the read data and the decoding process. These reliability data units (syndrome weights) serve as mediator information that guides the soft-decision decoder to make more accurate decisions about which bits are likely erroneous. The reliability data acts as a bridge that translates raw read data into informed decoding decisions, enhancing error correction capability
2Reliability
If soft-decision decoding is executed to improve error correction, then error correction reliability increases, but the processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary calculation of syndrome weights and reliability data units before the actual decoding operation. By pre-computing the reliability information for each bit position based on the read data and syndrome calculations, the system prepares optimized input data for the soft-decision decoder, which can then operate more efficiently with pre-processed reliability guidance rather than processing raw data from scratch
Solution Approach 2:
The patent segments the error correction process into distinct stages: reading multiple candidate data values with different biases, calculating syndrome weights for each candidate, determining reliability data units by comparing syndromes, and finally executing soft-decision decoding with the reliability-guided data. This segmentation allows each stage to be optimized independently, balancing computational complexity and processing time
3Measurement precision
If multiple read operations with different biases are performed to determine reliable data, then data accuracy improves, but the number of operations and time consumption increase
Solution Approach 1:
The patent performs partial redundant read operations by reading the same memory location multiple times with different read biases (e.g., different threshold voltages). Instead of reading once, it performs multiple reads with varying conditions to gather additional information about the true data value. This partial redundancy in reading operations provides excess information that is then processed to determine the most accurate data value and its reliability
Solution Approach 2:
The patent implements feedback by using the syndrome weights calculated from multiple read operations to inform the selection of the final read data. The syndrome calculation provides feedback about the consistency and reliability of each read candidate, and this feedback is used to select the most reliable data value for decoding. The system continuously refines its data selection based on syndrome feedback from multiple biased reads
Data Source
AI summary
A controller may include i) a storage memory configured to store N-bit read data and N reliability data units, ii) a decoder configured to execute a decoding operation for the read data based on the reliability data units, and iii) a processing circuit configured to determine a value of reliability data unit corresponding to I-th bit of read data based on an I-th bit of first read data, an I-th bit of second read data and a difference between first syndrome weight and second syndrome weight.