Abstract

Low-density parity check (LDPC) error correction decoders have become popular in diverse communications systems, owing to their strong error correction performance and their suitability to parallel hardware implementation. A great deal of research effort has been invested into the implementation of LDPC decoder designs on field-programmable gate array (FPGA) devices, in order to exploit their high processing speed, parallelism, and re-programmability. Meanwhile, a variety of application-specific integrated circuit implementations of multi-mode LDPC decoders exhibiting both inter-standard and intra-standard reconfiguration flexibility are available in the open literature. However, the high complexity of the adaptable routing and processing elements that are required by a flexible LDPC decoder has resulted in a lack of viable FPGA-based implementations. Hence in this paper, we propose a parameterisable FPGA-based LDPC decoder architecture, which supports run-time flexibility over any set of one or more quasi-cyclic LDPC codes. Additionally, we propose an off-line design flow, which may be used to automatically generate an optimized HDL description of our decoder, having support for a chosen selection of codes. Our implementation results show that the proposed architecture achieves a high level of design-time and run-time flexibility, whilst maintaining a reasonable processing throughput, hardware resource requirement, and error correction performance.

Highlights

  • A Flexible Field-Programmable Gate Array (FPGA)-Based Quasi-Cyclic LDPC DecoderAbstract—Low-Density Parity Check (LDPC) error correction decoders have become popular in diverse communications systems, owing to their strong error correction performance and their suitability to parallel hardware implementation

  • Low-Density Parity Check (LDPC) codes [1] constitute a class of Forward Error Correction (FEC) block codes that have been the focus of much research in the communications community over the past two decades

  • One of the many LDPC decoder parameters discussed in [7] is the level of processing parallelism, which is defined by the number of parallel Node Processing Units (NPUs) that are utilised to simultaneously process rows or columns of the Parity-Check Matrix (PCM)

Read more

Summary

A Flexible FPGA-Based Quasi-Cyclic LDPC Decoder

Abstract—Low-Density Parity Check (LDPC) error correction decoders have become popular in diverse communications systems, owing to their strong error correction performance and their suitability to parallel hardware implementation. A great deal of research effort has been invested into the implementation of LDPC decoder designs on Field-Programmable Gate Array (FPGA) devices, in order to exploit their high processing speed, parallelism and re-programmability. The high complexity of the adaptable routing and processing elements that are required by a flexible LDPC decoder has resulted in a lack of viable FPGA-based implementations. We propose a parameterisable FPGA-based LDPC decoder architecture, which supports run-time flexibility over any set of one or more Quasi-Cyclic (QC) LDPC codes. We propose an offline design flow, which may be used to automatically generate an optimised HDL description of our decoder, having support for a chosen selection of codes. Our implementation results show that the proposed architecture achieves a high level of design-time and run-time flexibility, whilst maintaining a reasonable processing throughput, hardware resource requirement and error correction performance. Tcp Clock cycles per P CNs tcb Clock cycles per Z CNs ti Clock cycles per iteration Ia Average number of iterations required

INTRODUCTION
BACKGROUND
LDPC codes
Background
Conclusion
Quasi-cyclic codes
LDPC decoder architectures
THE PROPOSED FLEXIBLE LDPC DECODER
Parallelism and decoding schedule
Memory organisation
Datapath
Programmable barrel shifters
NPU designs
Controller
PCM interpretation
HDL generation
NPU tree generation
IMPLEMENTATION RESULTS
Method
Decoders targeted at WiFi LDPC PCMs
Decoders targeted at PCMs from multiple families
CONCLUSIONS
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call