Permutation Sub-Matrix Checksum for Faster LDPC Decoding

Solution Overview

Problem

Existing error correction codes in flash memory, such as LDPC codes, face delays in decoding due to the need for post-iteration checksum computation, which can prolong decoding time and increase energy consumption.

Innovation Solution

A checksum structure is introduced that allows for 'on-the-fly' checksum computation, compatible with LDPC decoders, enabling real-time convergence detection and reducing decoding delays without additional complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If post-iteration checksum computation is performed after LDPC decoding, then data reliability is improved through error detection, but decoding time and energy consumption increase

Engineering Contradiction:

Improvedata reliabilityVSAvoiddecoding time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The checksum computation is performed preliminarily during the iterative decoding process rather than after completion. The syndrome computation and convergence detection are integrated into the decoding iterations, allowing the checksum to be ready when decoding completes, thus avoiding additional post-decoding computation time while maintaining error detection capability

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The checksum computation is merged with the LDPC decoding process by using the same syndrome computation mechanism. The convergence detection unit combines both ECC convergence detection and checksum computation into a single integrated process, eliminating redundant operations and reducing overall decoding time

Inventive Principle:

Principle #5Merging (Combining)

2Reliability

If post-iteration checksum computation is performed after LDPC decoding, then data reliability is improved through error detection, but energy consumption increases

Engineering Contradiction:

Improvedata reliabilityVSAvoidenergy consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The checksum computation is performed preliminarily during the iterative decoding process rather than after completion. The syndrome computation and convergence detection are integrated into the decoding iterations, allowing the checksum to be ready when decoding completes, thus avoiding additional post-decoding computation time and energy expenditure

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The checksum computation is merged with the LDPC decoding process by using the same syndrome computation mechanism. The convergence detection unit combines both ECC convergence detection and checksum computation into a single integrated process, eliminating redundant operations and reducing overall energy consumption

Inventive Principle:

Principle #5Merging (Combining)

3Reliability

If conventional CRC processing is performed on decoded codeword, then error detection capability is improved, but decoding delay increases

Engineering Contradiction:

Improveerror detection capabilityVSAvoiddecoding delay

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The conventional CRC processing step is extracted and replaced with an integrated convergence detection mechanism that performs both LDPC convergence detection and checksum computation simultaneously. This eliminates the sequential execution of CRC processing after decoding, removing the associated delay while maintaining error detection capability

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The checksum computation is merged with the LDPC decoding process by using the same syndrome computation mechanism. The convergence detection unit combines both ECC convergence detection and checksum computation into a single integrated process, eliminating redundant operations and reducing overall decoding time

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS8880987B2Checksum using sums of permutation sub-matrices

Publication Date: 2014.11.04 SANDISK TECHNOLOGIES LLC

AI summary

A method for encoding data bits includes computing checksum parity bits based on the data bits. A set of equations satisfied by the data bits and the checksum parity bits corresponds to a dense parity-check matrix. The dense parity-check matrix comprises sums of permutation sub-matrices.