Architectures de Codes Correcteurs d'Erreurs
Thèmes scientifiques :
- D - Télécommunications : compression, protection, transmission
Nous vous rappelons que, afin de garantir l'accès de tous les inscrits aux salles de réunion, l'inscription aux réunions est gratuite mais obligatoire.
15 personnes membres du GdR ISIS, et 25 personnes non membres du GdR, sont inscrits à cette réunion.
Capacité de la salle : 40 personnes.
Journée inter-GDR SoCSiP, ISIS : Architectures de Codes Correcteurs d'Erreurs.
4 novembre 2014, salle B007, Télécom Bretagne, BREST
« Architecture de codes correcteurs d’erreurs »
Journée inter GDR ISIS et SoCSiP
4 novembre 2014, salle B007, Télécom Bretagne, BREST
9h15 : accueil
9h45-10h15 : Stefan Weithoffer / Norbert Wehn (Technische Universität Kaiserslautern), Fully LTE-A compliant Turbo Decoder for Base Station Applications
10h15-10h45 : Li Meng (IMEC), ASIP LDPC for 802.11ad or ac standards.
10h45-11h15 : Benoît Parrein (Univ-Nantes), The Mojette Erasure Code
11h15-11h45 : Cafe Break
11h45:12h15 : David Declercq (ENSEA, Cergy Pontoise), Decoders LDPC FAID.
12h15-12h45 : Camille Leroux (IPB/ENSEIRB-MATMECA), Architecture of polar code decoder.
12h45-14h : Lunch
14h-14h30 : Valentin Savin (CEA LETI), Cost efficient FPGA implementations of Min-Sum and Self-Corrected-Min-Sum decoders.
14h30-15h : Philipp Schläfer / Norbert Wehn (Technische Universität Kaiserslautern), 100 Giga bits LDPC code
15h-15h30 : Charbel ABDEL NOUR (Télécom Bretagne), Title to be defined
15h30-15h50 : Cafe Break
15h50-16h20 : Emmanuel Boutillon (Lab-STICC/UBS), A Survey on Bit Flipping Algorithm
16h20-16h50 : Stefan Scholl / Norbert Wehn (Technische Universität Kaiserslautern), A High Speed Soft-RS Decoder
16h50 : Final discussion
Organisation pratique :
Les présentations seront diffusées en temps réel sur le site WEb :
Pour des raisons d’organisation (repas de midi), pouvez-vous vous inscrire à travers le GDR-ISIS ?
Résumés des contributions
Fully LTE-A compliant Turbo Decoder for Base Station Applications
Stefan Weithoffer / Norbert When (Technische Universität Kaiserslautern
Designing Turbo decoders for the latest LTE standard is a challenging task. We present a 1Gbit/s fully LTE-A compliant Turbo Decoder performing 14 half-iteration. Transport block CRCs are managed and a new latency reduced on the fly CRC calculation is proposed.
ASIP LDPC for 802.11ad or ac standards
Li Meng (IMEC)
The design of multi-Gbps LDPC decoder has become a hot topic in the recent years as the demand of the evaluation to 4G or even 5G network. IMEC designed a LDPC decoder processor. The ASIP core is targeting high throughput and it can be configured for the 802.11ad and the 802.11ac standards at the moment and is easy to be configured for the other standards.
The Mojette Erasure Code
Benoît Parrein (Univ-Nantes)
For twenty years, we developed at University of Nantes (IRCCyN lab) new algorithms and software for erasure coding based on discrete geometry and Radon transforms. Today, the solution is mature and allows incredible throughputs nearly close to the memcopy instruction. A comparison to Jerasure and Intel ISA libraries that implement optimized Reed-Solomon codes will be provided. An example of a complete integration to a scale-out NAS file system will conclude the talk.
Decoders LDPC FAID
David Declercq (ETIS, ENSEA)
Architecture of polar code decoder
Camille Leroux (IPB/ENSEIRB-MATMECA)
Polar coding has been recently proposed as a solution to achieve the capacity of binary input symetric memoryless channels. The inclusion of polar codes in future telecommunication standards highly depends on the efficiency of the hardware polar decoders. This presentation provides an overview of existing architectural solutions aiming at a more efficient implementation of the Successive Cancellation decoder. After a brief presentation of the SC decoding algorithm, we will highlight the different challenges of SC decoding implementation and present different architectural approaches that improve the efficiency of polar decoders.
Cost efficient FPGA implementations of Min-Sum and Self-Corrected-Min-Sum decoders
Valentin Savin (CEA LETI)
We present memory efficient FPGA implementations for layered quasi-cyclic (QC) LDPC decoders, based on the Self-Corrected Min-Sum (SCMS) algorithm. We address the problem of high memory overhead required by layered SCMS based decoders compared to conventional Min-Sum (MS), by proposing two improvements. These require changes in the flow/rule of the conventional SCMS algorithm, in order to avoid storing the signs and the erasure bits of the variable node messages. FPGA implementation results for WiMAX (1152, 2304) code show that the proposed architecture has a resource utilization with 17% less with respect to the one implementing conventional SCMS. Furthermore, it presents a similar cost to conventional MS, while having a 0.5 dB better error correction capability.
100 Giga bits LDPC code
Philipp Schläfer / Norbert Wehn (Technische Universität Kaiserslautern)
New standards have ever increasing data rates which are challenging to fulfil with state-of-the-art architectures. We present an efficient architecture optimised for very high throughput while maintaining good communications performance. At the same time we propose ways for increased energy efficiency in LDPC decoding.
Low-Complexity LDPC-coded Iterative MIMO Receiver Based on Belief Propagation algorithm for Detection
Charbel ABDEL NOUR (Lab-STICC/Télécom Bretagne)
A low-complexity Multiple Input Multiple Output receiver based on Belief Propagation (MIMO-BP) detector associated with a Non-Binary Low-Density Parity-Check (NB-LDPC) decoder is proposed in this talk. Such detection and decoding algorithms are represented thanks to a larger Joint Factor Graph (JFG). Shuffle schedule is also applied to efficiently exchange information between the detector and the decoder. Actions are undertaken at the detector, decoder and the iterative receiver levels in order to reduce overall computational complexity. An important reduction in terms of operations per iteration is obtained with a negligible performance penalty. Then, EXtrinsic Information Transfer (EXIT) charts are used to find a schedule in terms of number of iterations to be performed for which the proposed receiver achieves similar performance than a full-complexity iterative MIMO receiver.
A Survey on Bit Flipping Algorithm
Emmanuel Boutillon (Lab-STICC/UBS)
In this presentation, we will do a sort review on “single bit message”, with a focus on gradient bit flipping algorithm. We will show that this type of decoding algorithm can give performance close to the believe propagation algorithm with low hardware complexity.
A High Speed Soft-RS Decoder
Stefan Scholl / Norbert Wehn (Technische Universität Kaiserslautern)
Two ways to improve the error correction performance of today's channel coding systems are shown. First, a hardware architecture for soft decision decoding of Reed-Solomon codes is presented based on a information set decoding heuristic, that provides large gain in SNR at moderate complexity. Second, we present an efficient approach based on integer programming to perform exact ML decoding simulations.