Energy-drivenStatistical Profiling : DetectingSoftwareHotspots FayChang , KeithFarkas , andParthasarathyRanganathan CompaqWesternResearchLaboratories
semanticscholar（2002）
摘要
Energy is a critical resource in manycomputingsystems,motivatingthe needfor energy-efficient software design.Thisworkproposesa new approach,energy-drivenstatisticalprofiling, to helpsoftwaredevelopersreason abouttheenergy impactof software designdecisions.We describea prototypeimplementationof this approach for theItsypocketcomputingplatform.Our experimentalresultsusingtheprototypewith 13benchmarkprograms showthat (i) energy measur ementtoolsthat ignore system/kernel effectscangiveerroneousresultsaboutenergy hotspots,and(ii) usingexecutiontimeprofilesto developintuition aboutenergy usagecanoftenbemisleading.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要