{"title":"MultiView和Millipage——基于页面的dsm中的细粒度共享","authors":"Ayal Itzkovitz, A. Schuster","doi":"10.1145/296806.296830","DOIUrl":null,"url":null,"abstract":"In this paper we develop a novel technique, called MULTIVIEW, which enables implementation of page-based fine-grain DSMs. We show how the traditional techniques for implementing page-based DSMs can be extended to control the sharing granularity in a flexible way, even when the size of the sharing unit varies, and is smaller than the operating system's page size. The run-time overhead imposed in the proposed technique is negligible. We present a DSM system, called MILLIPAGE, which builds upon MULTIVIEW in order to support sharing in variable-size units. MILLIPAGE efficiently implements Sequential Consistency and shows comparable (sometimes superior) performance to related systems which use relaxed consistency models. It uses standard user-level operating system API and requires no compiler intervention, page twinning, diffs, code instrumentation, or sophisticated protocols. The resulting system is a thin software layer consisting mainly of a simple, clean protocol that handles page-faults.","PeriodicalId":90294,"journal":{"name":"Proceedings of the -- USENIX Symposium on Operating Systems Design and Implementation (OSDI). USENIX Symposium on Operating Systems Design and Implementation","volume":"10 11","pages":"215-228"},"PeriodicalIF":0.0000,"publicationDate":"1999-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"85","resultStr":"{\"title\":\"MultiView and Millipage — fine-grain sharing in page-based DSMs\",\"authors\":\"Ayal Itzkovitz, A. Schuster\",\"doi\":\"10.1145/296806.296830\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this paper we develop a novel technique, called MULTIVIEW, which enables implementation of page-based fine-grain DSMs. We show how the traditional techniques for implementing page-based DSMs can be extended to control the sharing granularity in a flexible way, even when the size of the sharing unit varies, and is smaller than the operating system's page size. The run-time overhead imposed in the proposed technique is negligible. We present a DSM system, called MILLIPAGE, which builds upon MULTIVIEW in order to support sharing in variable-size units. MILLIPAGE efficiently implements Sequential Consistency and shows comparable (sometimes superior) performance to related systems which use relaxed consistency models. It uses standard user-level operating system API and requires no compiler intervention, page twinning, diffs, code instrumentation, or sophisticated protocols. The resulting system is a thin software layer consisting mainly of a simple, clean protocol that handles page-faults.\",\"PeriodicalId\":90294,\"journal\":{\"name\":\"Proceedings of the -- USENIX Symposium on Operating Systems Design and Implementation (OSDI). USENIX Symposium on Operating Systems Design and Implementation\",\"volume\":\"10 11\",\"pages\":\"215-228\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1999-02-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"85\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the -- USENIX Symposium on Operating Systems Design and Implementation (OSDI). USENIX Symposium on Operating Systems Design and Implementation\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/296806.296830\",\"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 -- USENIX Symposium on Operating Systems Design and Implementation (OSDI). USENIX Symposium on Operating Systems Design and Implementation","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/296806.296830","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
MultiView and Millipage — fine-grain sharing in page-based DSMs
In this paper we develop a novel technique, called MULTIVIEW, which enables implementation of page-based fine-grain DSMs. We show how the traditional techniques for implementing page-based DSMs can be extended to control the sharing granularity in a flexible way, even when the size of the sharing unit varies, and is smaller than the operating system's page size. The run-time overhead imposed in the proposed technique is negligible. We present a DSM system, called MILLIPAGE, which builds upon MULTIVIEW in order to support sharing in variable-size units. MILLIPAGE efficiently implements Sequential Consistency and shows comparable (sometimes superior) performance to related systems which use relaxed consistency models. It uses standard user-level operating system API and requires no compiler intervention, page twinning, diffs, code instrumentation, or sophisticated protocols. The resulting system is a thin software layer consisting mainly of a simple, clean protocol that handles page-faults.