{"title":"Resource constrained dataflow retiming heuristics for VLIW ASIPs","authors":"M. Jacome, G. Veciana, C. Akturan","doi":"10.1109/HSC.1999.777383","DOIUrl":null,"url":null,"abstract":"This paper addresses issues in code generation of time critical loops for VLIW ASIPs with heterogenous distributed register structures. We discuss a code generation phasing whereby one first considers binding options that minimize the significant delays that may be incurred on such processors. Given such a binding we consider retiming, subject to code size constraints, so as to enhance performance. Finally a compatible schedule, minimizing latency, is sought. Our main focus in this paper is on the role retiming plays in this complex code generation problem. We propose heuristic algorithms for exploring code size/performance tradeoffs through retiming. Experimental results are presented indicating that the heuristics perform well on a sample of dataflows.","PeriodicalId":344739,"journal":{"name":"Proceedings of the Seventh International Workshop on Hardware/Software Codesign (CODES'99) (IEEE Cat. No.99TH8450)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1999-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"15","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the Seventh International Workshop on Hardware/Software Codesign (CODES'99) (IEEE Cat. No.99TH8450)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HSC.1999.777383","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 15
Abstract
This paper addresses issues in code generation of time critical loops for VLIW ASIPs with heterogenous distributed register structures. We discuss a code generation phasing whereby one first considers binding options that minimize the significant delays that may be incurred on such processors. Given such a binding we consider retiming, subject to code size constraints, so as to enhance performance. Finally a compatible schedule, minimizing latency, is sought. Our main focus in this paper is on the role retiming plays in this complex code generation problem. We propose heuristic algorithms for exploring code size/performance tradeoffs through retiming. Experimental results are presented indicating that the heuristics perform well on a sample of dataflows.