{"title":"Low power code generation of multiplication-free linear transforms","authors":"M. Mehendale, S. Sherlekar","doi":"10.1109/ICVD.1999.745122","DOIUrl":null,"url":null,"abstract":"The paper presents low power code generation of multiplication-free linear transforms targeted to both the register-rich RISC architectures and the single-register accumulator based DSP architectures. For register rich architectures, we present ordered chain-type DAC as the optimum structure for low power code generation of 1-dimensional transforms. For 2-dimensional transforms, we present an algorithm that performs instruction scheduling followed by register assignment for low power. For single-register architectures, we present a node re-ordering technique for reducing power dissipation. We present results to highlight the effectiveness of these techniques.","PeriodicalId":443373,"journal":{"name":"Proceedings Twelfth International Conference on VLSI Design. (Cat. No.PR00013)","volume":"118 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1999-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings Twelfth International Conference on VLSI Design. (Cat. No.PR00013)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICVD.1999.745122","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
The paper presents low power code generation of multiplication-free linear transforms targeted to both the register-rich RISC architectures and the single-register accumulator based DSP architectures. For register rich architectures, we present ordered chain-type DAC as the optimum structure for low power code generation of 1-dimensional transforms. For 2-dimensional transforms, we present an algorithm that performs instruction scheduling followed by register assignment for low power. For single-register architectures, we present a node re-ordering technique for reducing power dissipation. We present results to highlight the effectiveness of these techniques.