{"title":"用合作博弈论重新思考能源会计","authors":"Mian Dong, Tian Lan, Lin Zhong","doi":"10.1145/2639108.2639128","DOIUrl":null,"url":null,"abstract":"Energy accounting determines how much a software principal contributes to the total system energy consumption. It is the foundation for evaluating software and for operating system based energy management. While various energy accounting policies have been tried, there is no known way to evaluate them directly simply because it is hard to track all hardware usage by software in a heterogeneous multicore system like modern smartphones and tablets. In this work, we argue that energy accounting should be formulated as a cooperative game and that the Shapley value provides the ultimate ground truth for energy accounting policies. We reveal the important flaws of existing energy accounting policies based on the Shapley value theory and provide Shapley value-based energy accounting, a practical approximation of the Shapley value, for battery-powered mobile systems. We evaluate this approximation against existing energy accounting policies in two ways: (i) how well they identify the top energy consuming applications, and (ii) how effective they are in system energy management. Using a prototype based on Texas Instruments Pandaboard and smartphone workload, we experimentally demonstrate existing energy accounting policies can deviate by 400% in attributing energy consumption to running applications and can be up to 25% less effective in system energy management when compared to Shapley value-based energy accounting.","PeriodicalId":331897,"journal":{"name":"Proceedings of the 20th annual international conference on Mobile computing and networking","volume":"187 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2014-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"38","resultStr":"{\"title\":\"Rethink energy accounting with cooperative game theory\",\"authors\":\"Mian Dong, Tian Lan, Lin Zhong\",\"doi\":\"10.1145/2639108.2639128\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Energy accounting determines how much a software principal contributes to the total system energy consumption. It is the foundation for evaluating software and for operating system based energy management. While various energy accounting policies have been tried, there is no known way to evaluate them directly simply because it is hard to track all hardware usage by software in a heterogeneous multicore system like modern smartphones and tablets. In this work, we argue that energy accounting should be formulated as a cooperative game and that the Shapley value provides the ultimate ground truth for energy accounting policies. We reveal the important flaws of existing energy accounting policies based on the Shapley value theory and provide Shapley value-based energy accounting, a practical approximation of the Shapley value, for battery-powered mobile systems. We evaluate this approximation against existing energy accounting policies in two ways: (i) how well they identify the top energy consuming applications, and (ii) how effective they are in system energy management. Using a prototype based on Texas Instruments Pandaboard and smartphone workload, we experimentally demonstrate existing energy accounting policies can deviate by 400% in attributing energy consumption to running applications and can be up to 25% less effective in system energy management when compared to Shapley value-based energy accounting.\",\"PeriodicalId\":331897,\"journal\":{\"name\":\"Proceedings of the 20th annual international conference on Mobile computing and networking\",\"volume\":\"187 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2014-09-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"38\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 20th annual international conference on Mobile computing and networking\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/2639108.2639128\",\"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 20th annual international conference on Mobile computing and networking","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2639108.2639128","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Rethink energy accounting with cooperative game theory
Energy accounting determines how much a software principal contributes to the total system energy consumption. It is the foundation for evaluating software and for operating system based energy management. While various energy accounting policies have been tried, there is no known way to evaluate them directly simply because it is hard to track all hardware usage by software in a heterogeneous multicore system like modern smartphones and tablets. In this work, we argue that energy accounting should be formulated as a cooperative game and that the Shapley value provides the ultimate ground truth for energy accounting policies. We reveal the important flaws of existing energy accounting policies based on the Shapley value theory and provide Shapley value-based energy accounting, a practical approximation of the Shapley value, for battery-powered mobile systems. We evaluate this approximation against existing energy accounting policies in two ways: (i) how well they identify the top energy consuming applications, and (ii) how effective they are in system energy management. Using a prototype based on Texas Instruments Pandaboard and smartphone workload, we experimentally demonstrate existing energy accounting policies can deviate by 400% in attributing energy consumption to running applications and can be up to 25% less effective in system energy management when compared to Shapley value-based energy accounting.