{"title":"基于遗传算法的领域可重构阵列引擎","authors":"Wing On Fung, T. Arslan, S. Khawam","doi":"10.1109/AHS.2006.48","DOIUrl":null,"url":null,"abstract":"Domain-specific reconfigurable arrays have shown to provide an efficient trade-off between flexibility of FPGA and performance of ASIC circuit. Nonetheless, the design of these heterogeneous arrays is a labour intensive process. Furthermore, the manual creation of the array architecture could not have been fully optimised, hence limiting their performance. This paper presents a placement technique for mapping logic elements into heterogeneous reconfigurable arrays. At its core, it implements a genetic algorithm, which was used to reduce the span of all the interconnections as well as critical delay. It therefore minimises the amount of routing resource required in the architecture. The algorithm was tested on two arrays implementing DCT and speech coding. The resulting architecture achieves optimal close to that of an expert designer in a fraction of the time","PeriodicalId":232693,"journal":{"name":"First NASA/ESA Conference on Adaptive Hardware and Systems (AHS'06)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2006-06-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":"{\"title\":\"Genetic Algorithm based Engine for Domain-Specific Reconfigurable Arrays\",\"authors\":\"Wing On Fung, T. Arslan, S. Khawam\",\"doi\":\"10.1109/AHS.2006.48\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Domain-specific reconfigurable arrays have shown to provide an efficient trade-off between flexibility of FPGA and performance of ASIC circuit. Nonetheless, the design of these heterogeneous arrays is a labour intensive process. Furthermore, the manual creation of the array architecture could not have been fully optimised, hence limiting their performance. This paper presents a placement technique for mapping logic elements into heterogeneous reconfigurable arrays. At its core, it implements a genetic algorithm, which was used to reduce the span of all the interconnections as well as critical delay. It therefore minimises the amount of routing resource required in the architecture. The algorithm was tested on two arrays implementing DCT and speech coding. The resulting architecture achieves optimal close to that of an expert designer in a fraction of the time\",\"PeriodicalId\":232693,\"journal\":{\"name\":\"First NASA/ESA Conference on Adaptive Hardware and Systems (AHS'06)\",\"volume\":\"75 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2006-06-15\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"5\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"First NASA/ESA Conference on Adaptive Hardware and Systems (AHS'06)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/AHS.2006.48\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"First NASA/ESA Conference on Adaptive Hardware and Systems (AHS'06)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AHS.2006.48","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Genetic Algorithm based Engine for Domain-Specific Reconfigurable Arrays
Domain-specific reconfigurable arrays have shown to provide an efficient trade-off between flexibility of FPGA and performance of ASIC circuit. Nonetheless, the design of these heterogeneous arrays is a labour intensive process. Furthermore, the manual creation of the array architecture could not have been fully optimised, hence limiting their performance. This paper presents a placement technique for mapping logic elements into heterogeneous reconfigurable arrays. At its core, it implements a genetic algorithm, which was used to reduce the span of all the interconnections as well as critical delay. It therefore minimises the amount of routing resource required in the architecture. The algorithm was tested on two arrays implementing DCT and speech coding. The resulting architecture achieves optimal close to that of an expert designer in a fraction of the time