Connecting campuses to promote excellence in math and science


David Bunde

Knox College
Computer Science
Power-aware scheduling
Power consumption has always been a major issue for users of laptops and other mobile devices, but its importance has increased over time because processor power consumption is growing faster than battery capacity. To cope with this problem, processors for some mobile devices now incorporate speed scaling, which allows the system to tradeoff performance and battery life by slowing the clock speed. An important question is how to make this tradeoff to achieve the best possible performance given a battery's limited power supply. This talk discusses scheduling with speed scaling under two common measures of schedule quality, completion time of the last job and response time. Both these scheduling problems have highly-structured optimal solutions, but one turns out to have a simple algorithm while the other is impossible to solve exactly.
Apply for monies for this speaker