Pipelined LDPC Layer Decoder for Overlapped Circulant Dependencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
LDPC layer decoders face delays due to dependencies between circulants in neighboring layers, which hinder efficient data decoding in systems with overlapped circulants, leading to potential errors and reduced processing speed.
Innovation Solution
Implementing a pipelined LDPC layer decoder that uses data from earlier decoding iterations for overlapped circulants, with increased memory to store calculated data and flags to indicate overlapped circulants, allowing out-of-order circulant processing to avoid waiting for recent results, and using ROM to control data usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data from the most recent iteration is used for decoding, then decoding accuracy is improved, but processing delays occur due to dependencies between overlapped circulants
Solution Approach 1:
The patent stores data from previous iterations in memory before it is needed for decoding overlapped circulants. By preparing and storing this data in advance (in the second memory structure), the decoder can retrieve it immediately when processing overlapped circulants, avoiding delays while maintaining decoding accuracy.
Solution Approach 2:
The patent introduces a dual-memory system where the second memory structure acts as an intermediary, holding copies of data from previous iterations. This intermediary storage allows the decoding process to access historical data without waiting for it to be regenerated, effectively mediating between the need for accurate data and the timing requirements of parallel processing.
2Reliability
If processing waits for recent circulant results, then decoding correctness is maintained, but processing speed decreases
Solution Approach 1:
Data from previous iterations is pre-calculated and stored in the second memory structure before being needed. This preliminary action allows the decoder to maintain correctness by using accurate historical data while achieving higher processing speeds through parallel pipelined operation without waiting for recent results to be computed.
Solution Approach 2:
The patent transitions from sequential processing to a pipelined parallel processing architecture. By organizing memories and processing units in parallel dimensions, the system can process multiple circulants simultaneously, using data from previous iterations stored in the second memory structure, thereby increasing productivity without sacrificing correctness.
3Productivity
If memory is increased to store earlier iteration data, then continuous decoding is enabled, but device complexity increases
Solution Approach 1:
The patent divides the memory system into two distinct memory structures: a first memory for current iteration data and a second memory for previous iteration data. This segmentation allows the system to manage complexity by organizing storage functions separately, enabling continuous decoding through pipelined processing while keeping each memory structure's design relatively simple and modular.
Solution Approach 2:
The dual-memory system serves multiple functions: the first memory holds current iteration data for active processing, while the second memory stores historical data for overlapped circulant decoding. This multi-functionality allows the increased memory capacity to support both continuous pipelined operation and accurate decoding of overlapped circulants, justifying the complexity increase through enhanced productivity.
Data Source
AI summary
The present inventions are related to systems and methods for decoding data in an LDPC layer decoder for LDPC codes with overlapped circulants.