{"title":"A novel scheme to generate optimal memcpy assembly code","authors":"Huan Ying, Hao Zhu, Donghui Wang, C. Hou","doi":"10.1109/ICIST.2013.6747619","DOIUrl":null,"url":null,"abstract":"Memcpy function is invoked frequently by digital signal processor to handle data transfer during encoding and decoding of multimedia. This paper presents a novel scheme to generate optimal memcpy assembly code for a reduced instruction set computer architecture. Based on the ordered binary decision diagram and the concept of data mining, the proposed solution completely exploits the available optimization according to the input parameters of memcpy. Thus, the memcpy function assembly code generated by our proposal accomplishes the directed-optimal data transfer during execution. Experimental results showed that the solution improves the performance of memcpy effectively. It reduced 56.5c/o executed cycles on average. Moreover, the approach of generation for memcpy at assembly level is portable for any other reduced instruction set computer architecture with modification of the customization file.","PeriodicalId":415759,"journal":{"name":"2013 IEEE Third International Conference on Information Science and Technology (ICIST)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2013-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2013 IEEE Third International Conference on Information Science and Technology (ICIST)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICIST.2013.6747619","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Memcpy function is invoked frequently by digital signal processor to handle data transfer during encoding and decoding of multimedia. This paper presents a novel scheme to generate optimal memcpy assembly code for a reduced instruction set computer architecture. Based on the ordered binary decision diagram and the concept of data mining, the proposed solution completely exploits the available optimization according to the input parameters of memcpy. Thus, the memcpy function assembly code generated by our proposal accomplishes the directed-optimal data transfer during execution. Experimental results showed that the solution improves the performance of memcpy effectively. It reduced 56.5c/o executed cycles on average. Moreover, the approach of generation for memcpy at assembly level is portable for any other reduced instruction set computer architecture with modification of the customization file.