{"title":"Towards compilation of streaming programs into FPGA hardware","authors":"F. Plavec, Z. Vranesic, S. Brown","doi":"10.1109/FDL.2008.4641423","DOIUrl":null,"url":null,"abstract":"There is an increasing need for automated conversion of high-level design descriptions into hardware. We present a flow that converts a software application written in the Brook streaming language into a hardware description targeting FPGAs. We use a combination of our source-to-source compiler and a commercial C2H behavioral synthesis compiler. Our approach results in a significant through-put increase compared to software and ordinary C2H results (up to 8.9X and 4.3X, respectively). The throughput can be further increased by using more hardware resources to exploit data parallelism available in streaming applications.","PeriodicalId":202172,"journal":{"name":"2008 Forum on Specification, Verification and Design Languages","volume":"7 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2008-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"20","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2008 Forum on Specification, Verification and Design Languages","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/FDL.2008.4641423","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 20
Abstract
There is an increasing need for automated conversion of high-level design descriptions into hardware. We present a flow that converts a software application written in the Brook streaming language into a hardware description targeting FPGAs. We use a combination of our source-to-source compiler and a commercial C2H behavioral synthesis compiler. Our approach results in a significant through-put increase compared to software and ordinary C2H results (up to 8.9X and 4.3X, respectively). The throughput can be further increased by using more hardware resources to exploit data parallelism available in streaming applications.