{"title":"The Gauss machine: A Galois-enhanced quadratic residue number system systolic array","authors":"J. Mellott, Jermy C. Smith, F. Taylor","doi":"10.1109/ARITH.1993.378097","DOIUrl":null,"url":null,"abstract":"The Gauss machine is a SIMD systolic array architecture that takes advantage of the Galois-enhanced residue number system (GEQRNS) to form reduced-complexity arithmetic elements. The Gauss machine is targeted at front-end signal and image processing applications. A discrete prototype that achieves a peak rating of 320 million complex arithmetic operations per second while operating at 10 MHz has been constructed. A VLSI implementation of the Gauss machine's processor cell has been created. The VLSI implementation is implemented in 2.0-/spl mu/m CMOS and achieves greater than 20-MHz performance, using less than 2.0-mm/sup 2/ die area. It is shown that techniques for defect tolerance in RNS systolic arrays can result in substantial yield enhancement, thereby making larger than conventional (ULSI) systems possible.<<ETX>>","PeriodicalId":414758,"journal":{"name":"Proceedings of IEEE 11th Symposium on Computer Arithmetic","volume":"5 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1993-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of IEEE 11th Symposium on Computer Arithmetic","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ARITH.1993.378097","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 11
Abstract
The Gauss machine is a SIMD systolic array architecture that takes advantage of the Galois-enhanced residue number system (GEQRNS) to form reduced-complexity arithmetic elements. The Gauss machine is targeted at front-end signal and image processing applications. A discrete prototype that achieves a peak rating of 320 million complex arithmetic operations per second while operating at 10 MHz has been constructed. A VLSI implementation of the Gauss machine's processor cell has been created. The VLSI implementation is implemented in 2.0-/spl mu/m CMOS and achieves greater than 20-MHz performance, using less than 2.0-mm/sup 2/ die area. It is shown that techniques for defect tolerance in RNS systolic arrays can result in substantial yield enhancement, thereby making larger than conventional (ULSI) systems possible.<>