{"title":"A flexible ASIP decoder for combined binary and non-binary LDPC codes","authors":"F. Naessens, A. Bourdoux, A. Dejonghe","doi":"10.1109/SCVT.2010.5720462","DOIUrl":null,"url":null,"abstract":"This paper describes the implementation of a flexible combined binary and non-binary LDPC decoder. The ASIP architecture can be configured allowing re-use between both modes. Key in the architecture is parallelization, which is exploited in the SIMD engine. Binary LDPC codes intrinsically enables parallelization through layered decoding while in the non-binary case different trade-offs can be made. The implementation choice was made base on minimal memory requirement and computational effort. For a combination of supporting binary LDPC present within WLAN and WiMAX standard with non-binary GF(8) LDPC codes, a total area of 5.4 sqmm in commercial 65nm technology would be required. This size can be reduced towards 3.4 sqmm if only half of the non-binary decoding throughput is required.","PeriodicalId":344975,"journal":{"name":"2010 17th IEEE Symposium on Communications and Vehicular Technology in the Benelux (SCVT2010)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2010-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"9","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2010 17th IEEE Symposium on Communications and Vehicular Technology in the Benelux (SCVT2010)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SCVT.2010.5720462","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 9
Abstract
This paper describes the implementation of a flexible combined binary and non-binary LDPC decoder. The ASIP architecture can be configured allowing re-use between both modes. Key in the architecture is parallelization, which is exploited in the SIMD engine. Binary LDPC codes intrinsically enables parallelization through layered decoding while in the non-binary case different trade-offs can be made. The implementation choice was made base on minimal memory requirement and computational effort. For a combination of supporting binary LDPC present within WLAN and WiMAX standard with non-binary GF(8) LDPC codes, a total area of 5.4 sqmm in commercial 65nm technology would be required. This size can be reduced towards 3.4 sqmm if only half of the non-binary decoding throughput is required.