Iteration Overlap for Low-Latency Turbo Decoding - Equipe Channel and source coding solutions for emerging digital communication and storage systems Access content directly
Conference Papers Year : 2023

Iteration Overlap for Low-Latency Turbo Decoding

Abstract

Achieving high decoding throughput and latency has been challenging for turbo decoders due to the limitations in terms of parallelism on component decoder level. To alleviate this issue, we propose an iteration overlap scheme able to apply a decoding schedule tailored to both, the decoder hardware architecture and the interleaver constraints. The proposal aims to minimize the achieved decoding latency without penalizing performance when compared to baseline decoders. To that end, we formulate the window schedule optimization problem when applying iteration overlap in pipelined Turbo Decoder hardware architectures. Then, we propose a method to find optimal window schedules under realistic assumptions. Results demonstrate that latency is reduced by 20 − 25% for most Long Term Evolution (LTE) interleaver configurations. For specific interleavers, the achieved latency reduction can be as high as 62%. This indicates that further latency savings could be achieved if iteration overlap is taken into account when designing interleavers.
Fichier principal
Vignette du fichier
ISTC_2023_Overlap_Interleaver.pdf (434.47 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-04168135 , version 1 (21-07-2023)

Identifiers

  • HAL Id : hal-04168135 , version 1

Cite

Stefan Weithoffer, Ghazi Aousaji, Jérémy Nadal, Charbel Abdel Nour. Iteration Overlap for Low-Latency Turbo Decoding. ISTC 2023 : 12th International Symposium on Topics in Coding, Sep 2023, Brest, France. ⟨hal-04168135⟩
25 View
31 Download

Share

Gmail Facebook X LinkedIn More