Minimizing total flow time and total completion time with immediate dispatching

Avrahami, Nir; Azar, Yossi

doi:10.1145/777412.777415

Cited by 50 publications

(63 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, for scheduling non-batched sequential or parallel jobs, some researchers [4,6,11,16,20,26,34] have presented various algorithms and analyzed their performances. In particular, using resource augmentation analysis [26], which allows an online algorithm to have access to more processors than the optimal, Edmonds [16] has shown that EQUI achieves O(1)-speed O(1)-competitiveness with respect to the mean response time on parallel jobs modeled by multiple phases of arbitrary nondecreasing and sub-linear speedup functions.…”

Section: Discussionmentioning

confidence: 99%

Improved results for scheduling batched parallel jobs by using a generalized analysis framework

Sun

Hsu

2010

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Improved results for scheduling batched parallel jobs by using a generalized analysis framework

Sun

Hsu

2010

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

“…In the past, when energy was not a major concern, the objective of scheduling algorithms was to minimize the total response time (also called flow time) of all tasks where processors were running at fixed speeds (e.g [20,1]). The response time is the time elapsed since a task arrives until it is completed.…”

Section: Related Workmentioning

confidence: 99%

“…• At any given time, the th j stream processor processes a task stored in the first index ( ) j , 1 of the memory queue.…”

Section: ≤ ≤mentioning

confidence: 99%

Cost Minimization for Scheduling Parallel, Single-Threaded, Heterogeneous, Speed-Scalable Processors

Khogali

Das

2013

2013 International Conference on Parallel and Distributed Systems

View full text Add to dashboard Cite

We introduce an online scheduling algorithm to optimally assign a set of arriving heterogeneous tasks to heterogeneous speed-scalable processors. The goal of our algorithm is to minimize the total cost of response time and energy consumption (TCRTEC) of the tasks. We have three contributions that constitute the algorithm. First, we propose a novel task dispatching strategy for assigning the tasks to the processors. Second, we propose a novel preemptive service discipline called Smallest remaining Computation Volume Per unit Price of response Time (SCVPPT) to schedule the tasks on the assigned processor. Third, we propose a dynamic speedscaling function that explicitly determines the optimum processing rate of each task. In our work, the processors are heterogeneous in that they may differ in their hardware specifications with respect to maximum processing rate and power functions. Tasks are heterogeneous in terms of computation volume and processing requirements. We also consider that the unit price of response time for each task is heterogeneous. Each task's unit price of response time is allowed to differ because the user may be willing to pay higher/lower unit prices for certain tasks; thereby increasing/decreasing their optimum processing rates. In our SCVPPT discipline, a task's scheduling priority is influenced by its remaining computation volume as well as its unit price of response time. Our simulation results show that SCVPPT outperforms the two known service disciplines, Shortest Remaining Processing Time (SRPT) and the First Come First Serve (FCFS), in terms of minimizing the TCRTEC performance metric. The results also show that the algorithm's dispatcher outperforms the well known Round Robin dispatcher when the processors are heterogeneous. We focus on multi-buffer, single-threading where a set of tasks is allocated to a given processor, but only one task is processed at a time until completion unless preemption is dictated by the service discipline.

show abstract

“…While all these algorithms do not migrate jobs, they hold jobs in the pool between time of release and the assignment to a machine. Avrahami and Azar [1] developed an algorithm without migration and with immediate dispatch (each job is assigned to a machine upon its release) that achieves the same performance of the algorithm with delayed dispatch of [2].…”

Section: Afterwordmentioning

confidence: 99%

Approximating total flow time on parallel machines

Leonardi

Raz

2007

Journal of Computer and System Sciences

View full text Add to dashboard Cite

We consider the problem of optimizing the total flow time of a stream of jobs that are released over time in a multiprocessor setting. This problem is NP-hard even when there are only two machines and preemption is allowed. Although the total (or average) flow time is widely accepted as a good measurement of the overall quality of service, no approximation algorithms were known for this basic scheduling problem. This paper contains two main results. We first prove that when preemption is allowed, Shortest Remaining Processing Time (SRPT) is an O(log(min{n/m, P})) approximation algorithm for the total flow time, where it is the number of jobs, In is the number of machines, and P is the ratio between the maximum and the minimum processing time of a job. We also provide an R(log( + P)) lower bound on the (worst case) competitive ratio of any randomized algorithm for the on-line problem in which jobs are known at their release times. Thus, we show that up to a constant factor SRPT is an optimal on-line algorithm. Our second main result addresses the non-preemptive case. We present a general technique that allows to transform any preemptive solution into a non-preemptive solution at the expense of an O(root n/m) factor in the approximation ratio of the total flow time. Combining this technique with our previous result yields an O(root n/m log n/m) approximation algorithm for this case. We also show an Omega (n(1/3-is an element of)) lower bound on the approximabiiity of this problem (assuming P not equal NP). (c) 2006 Elsevier Inc. All rights reserved

show abstract

Minimizing total flow time and total completion time with immediate dispatching

Cited by 50 publications

References 17 publications

Improved results for scheduling batched parallel jobs by using a generalized analysis framework

Improved results for scheduling batched parallel jobs by using a generalized analysis framework

Cost Minimization for Scheduling Parallel, Single-Threaded, Heterogeneous, Speed-Scalable Processors

Approximating total flow time on parallel machines

Contact Info

Product

Resources

About