Speed Scaling for Weighted Flow Time

Abstract

Abstract. Intel's SpeedStep and AMD's PowerNOW technologies allow the Windows XP operating system to dynamically change the speed of the processor to prolong battery life. In this setting, the operating system must not only have a job selection policy to determine which job to run, but also a speed scaling policy to determine the speed at which the job will be run. We give an online speed scaling algorithm that is O(1)-competitive for the objective of weighted flow time plus energy. This algorithm also allows us to efficiently construct an O(1)-approximate schedule for minimizing weighted flow time subject to an energy constraint.

Nonclairvoyantly scheduling power-heterogeneous processors

“…The competitive ratio was improved by [LLTW08b] for the unweighted case using a potential function specifically tailored to integer flow. [BCLL08] extended the results of [BPS09] to the bounded speed model, and [CEL + 09] gave a nonclairvoyant algorithm that is O(1)-competitive.…”

“…[LLTW08a] show that an algorithm that uses a variation of round robin for the assignment policy, and uses the job selection and speed scaling policies from [BPS09], is scalable for this problem. [CEP09] show that bounded-competitiveness for the objective of flow plus energy is not achievable on multiprocessors if jobs can be run simultaneously on multiprocessors, and have varying speed-ups (i.e jobs have different degrees of parallelism).…”

Optimal energy trade-off schedules

“…[4], [5] showed that a natural online algorithm is 2-competitive for the objective of a linear combination of energy and total (unweighted) (integer) delay. Previously, [3], [6], [7], [8], [9], [10], [11] gave online algorithms with competitive analyses in the case that the power function was of the form s α . [12] first introduced the energy-delay product as a metric, and its use has been prevalent since then.…”

Energy‐aware scheduling of bag‐of‐tasks applications on master–worker platforms

“…Contrarily to [14] where tasks are gathered into blocks and scheduled with increasing speed in order to minimize the makespan, here the authors prove that the speed of the blocks needs to be decreasing in order to minimize both total flow time and energy consumption. Bansal et al [16] improved a part of this work and gave algorithms for the more general problem with arbitrary jobs.…”