{"title":"一种利用寄存器存在图的寄存器分配技术","authors":"A. Koseki, Y. Fukazawa, H. Komatsu","doi":"10.1109/ICPP.1997.622673","DOIUrl":null,"url":null,"abstract":"Optimizing compilation is very important for generating code sequences in order to utilize the characteristics of processor architectures. One of the most essential optimization techniques is register allocation. In register allocation that takes account of instruction-level parallelism, anti-dependences generated when the same register is allocated to different variables, and spill code generated when the number of registers is insufficient should be handled in such a way that the parallelism in a program is not lost. In our method, we realized register allocation using a new data structure called the register existence graph, in which the parallelism in program is well expressed.","PeriodicalId":221761,"journal":{"name":"Proceedings of the 1997 International Conference on Parallel Processing (Cat. No.97TB100162)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1997-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A register allocation technique using register existence graph\",\"authors\":\"A. Koseki, Y. Fukazawa, H. Komatsu\",\"doi\":\"10.1109/ICPP.1997.622673\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Optimizing compilation is very important for generating code sequences in order to utilize the characteristics of processor architectures. One of the most essential optimization techniques is register allocation. In register allocation that takes account of instruction-level parallelism, anti-dependences generated when the same register is allocated to different variables, and spill code generated when the number of registers is insufficient should be handled in such a way that the parallelism in a program is not lost. In our method, we realized register allocation using a new data structure called the register existence graph, in which the parallelism in program is well expressed.\",\"PeriodicalId\":221761,\"journal\":{\"name\":\"Proceedings of the 1997 International Conference on Parallel Processing (Cat. No.97TB100162)\",\"volume\":\"38 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1997-08-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 1997 International Conference on Parallel Processing (Cat. No.97TB100162)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICPP.1997.622673\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 1997 International Conference on Parallel Processing (Cat. No.97TB100162)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICPP.1997.622673","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A register allocation technique using register existence graph
Optimizing compilation is very important for generating code sequences in order to utilize the characteristics of processor architectures. One of the most essential optimization techniques is register allocation. In register allocation that takes account of instruction-level parallelism, anti-dependences generated when the same register is allocated to different variables, and spill code generated when the number of registers is insufficient should be handled in such a way that the parallelism in a program is not lost. In our method, we realized register allocation using a new data structure called the register existence graph, in which the parallelism in program is well expressed.