Lap-Kei Lee

Scheduling for Speed Bounded Processors

We consider online scheduling algorithms in the dynamic speedscaling model, where a processor can scale its speed between 0 andsome maximum speed T. The processor uses energy at ratesαwhen run at speed s,where α> 1 is a constant. Most modern processorsuse dynamic speed scaling to manage their energy usage. This leadsto the problem of designing execution strategies that are bothenergy efficient, and yet have almost optimum performance. We consider two problems in this model and give essentiallyoptimum possible algorithms for them. In the first problem, jobswith arbitrary sizes and deadlines arrive online and the goal is tomaximize the throughput, i.e. the total size of jobs completedsuccessfully. We give an algorithm that is 4-competitive forthroughput and O(1)-competitive for the energy used. Thisimproves upon the 14 throughput competitive algorithm of Chan etal. [10]. Our throughput guarantee is optimal as any onlinealgorithm must be at least 4-competitive even if the energy concernis ignored [7]. In the second problem, we consider optimizing thetrade-off between the total flow time incurred and the energyconsumed by the jobs. We give a 4-competitive algorithm to minimizetotal flow time plus energy for unweighted unit size jobs, and a (2+ o(1)) α/ln α-competitivealgorithm to minimize fractional weighted flow time plus energy.Prior to our work, these guarantees were known only when theprocessor speed was unbounded (T= ∞) [4].

A simpler and more efficient deterministic scheme for finding frequent items over sliding windows

In this paper, we give a simple scheme for identifying ε-approximate frequent items over a sliding window of size <i>n</i>. Our scheme is deterministic and does not make any assumption on the distribution of the item frequencies. It supports <i>O</i>(1/ε) update and query time, and uses <i>O</i>(1/ε) space. It is very simple; its main data structures are just a few short queues whose entries store the position of some items in the sliding window. We also extend our scheme for variable-size window. This extended scheme uses <i>O</i>(1/ε log(ε<i>n</i>)) space.

SOAP3-dp: Fast, Accurate and Sensitive GPU-Based Short Read Aligner

To tackle the exponentially increasing throughput of Next-Generation Sequencing (NGS), most of the existing short-read aligners can be configured to favor speed in trade of accuracy and sensitivity. SOAP3-dp, through leveraging the computational power of both CPU and GPU with optimized algorithms, delivers high speed and sensitivity simultaneously. Compared with widely adopted aligners including BWA, Bowtie2, SeqAlto, CUSHAW2, GEM and GPU-based aligners BarraCUDA and CUSHAW, SOAP3-dp was found to be two to tens of times faster, while maintaining the highest sensitivity and lowest false discovery rate (FDR) on Illumina reads with different lengths. Transcending its predecessor SOAP3, which does not allow gapped alignment, SOAP3-dp by default tolerates alignment similarity as low as 60%. Real data evaluation using human genome demonstrates SOAP3-dps power to enable more authentic variants and longer Indels to be discovered. Fosmid sequencing shows a 9.1% FDR on newly discovered deletions. SOAP3-dp natively supports BAM file format and provides the same scoring scheme as BWA, which enables it to be integrated into existing analysis pipelines. SOAP3-dp has been deployed on Amazon-EC2, NIH-Biowulf and Tianhe-1A.

Speed Scaling Functions for Flow Time Scheduling Based on Active Job Count

We study online scheduling to minimize flow time plus energy usage in the dynamic speed scaling model. We devise new speed scaling functions that depend on the number of active jobs, replacing the existing speed scaling functions in the literature that depend on the remaining work of active jobs. The new speed functions are more stable and also more efficient. They can support better job selection strategies to improve the competitive ratios of existing algorithms [8,5], and, more importantly, to remove the requirement of extra speed. These functions further distinguish themselves from others as they can readily be used in the non-clairvoyant model (where the size of a job is only known when the job finishes). As a first step, we study the scheduling of batched jobs (i.e., jobs with the same release time) in the non-clairvoyant model and present the first competitive algorithm for minimizing flow time plus energy (as well as for weighted flow time plus energy); the performance is close to optimal.

Optimizing throughput and energy in online deadline scheduling

This article extends the study of online algorithms for energy-efficient deadline scheduling to the overloaded setting. Specifically, we consider a processor that can vary its speed between 0 and a maximum speed T to minimize its energy usage (the rate is believed to be a cubic function of the speed). As the speed is upper bounded, the processor may be overloaded with jobs and no scheduling algorithms can guarantee to meet the deadlines of all jobs. An optimal schedule is expected to maximize the throughput, and furthermore, its energy usage should be the smallest among all schedules that achieve the maximum throughput. In designing a scheduling algorithm, one has to face the dilemma of selecting more jobs and being conservative in energy usage. If we ignore energy usage, the best possible online algorithm is 4-competitive on throughput [Koren and Shasha 1995]. On the other hand, existing work on energy-efficient scheduling focuses on a setting where the processor speed is unbounded and the concern is on minimizing the energy to complete all jobs; O(1)-competitive online algorithms with respect to energy usage have been known [Yao et al. 1995; Bansal et al. 2007a; Li et al. 2006]. This article presents the first online algorithm for the more realistic setting where processor speed is bounded and the system may be overloaded; the algorithm is O(1)-competitive on both throughput and energy usage. If the maximum speed of the online scheduler is relaxed slightly to (1+&epsis;)T for some &epsis; > 0, we can improve the competitive ratio on throughput to arbitrarily close to one, while maintaining O(1)-competitiveness on energy usage.

Nonclairvoyant Speed Scaling for Flow and Energy

We study online nonclairvoyant speed scaling to minimize total flow time plus energy. We first consider the traditional model where the power function is

Energy efficient deadline scheduling in two processor systems

P(s)=s^\alpha

Non-clairvoyant speed scaling for weighted flow time

. We give a nonclairvoyant algorithm that is shown to be

Continuous Monitoring of Distributed Data Streams over a Time-Based Sliding Window

O(\alpha^3)

Sleep with Guilt and Work Faster to Minimize Flow Plus Energy

-competitive. We then show an