Proceedings Article10.1145/331532.331580
Stochastic Scheduling
Jennifer M. Schopf,Francine Berman +1 more
- 01 Jan 1999
136
TL;DR: It is demonstrated that a stochastic scheduling policy based on time-balancing for data parallel applications whose execution behavior can be represented as a normal distribution can achieve good and predictable performance for the application as evaluated by several performance measures.
read more
Abstract: There is a current need for scheduling policies that can leverage the performance variability of resources on multi-user clusters. We develop one solution to this problem called stochastic scheduling that utilizes a distribution of application execution performance on the target resources to determine a performance-efficient schedule. In this paper, we define a stochastic scheduling policy based on time-balancing for data parallel applications whose execution behavior can be represented as a normal distribution. Using three distributed applications on two contended platforms, we demonstrate that a stochastic scheduling policy can achieve good and predictable performance for the application as evaluated by several performance measures.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Scheduling Algorithms for Grid Computing: State of the Art and Open Problems
Fangpeng Dong,Selim G. Akl +1 more
- 01 Jan 2006
TL;DR: This survey provides a review of the subject of Grid scheduling mainly from the perspective of scheduling algorithms, and identifies the challenges and state of the art of current research.
Robust scheduling of a two-machine flow shop with uncertain processing times
TL;DR: A measure of schedule robustness that explicitly considers the risk of poor system performance over all potential realizations of job processing times is presented and results indicate that robust schedules provide effective hedges against processing time uncertainty while maintaining excellent expected makespan performance.
230
Online prediction of the running time of tasks
Peter A. Dinda
- 01 Jun 2001
TL;DR: The Running Time Advisor is described and evaluated, a system that can predict the running time of a compute-bound task on a typical shared, unreserved commodity host from linear time series predictions of host load and takes the form of a confidence interval that neatly expresses error associated with the measurement and prediction processes.
Online Prediction of the Running Time of Tasks
TL;DR: The Running Time Advisor is described and evaluated, a system that can predict the running time of a compute-bound task on a typical shared, unreserved commodity host from linear time series predictions of host load and takes the form of a confidence interval that neatly expresses error associated with the measurement and prediction processes.
107
Adaptive Performance Prediction for Distributed Data-Intensive Applications
M. Faerman,Alan Su,Richard Wolski,Francine Berman +3 more
- 01 Jan 1999
TL;DR: This paper introduces a performance prediction method, AdRM (Adaptive Regression Modeling), to determine file transfer times for network-bound distributed data-intensive applications, and demonstrates the effectiveness of the method on two distributed data applications.
References
The traveling salesman problem
Eugene L. Lawler,Jan Karel Lenstra,Alexander H. G. Rinnooy Kan,David B. Shmoys +3 more
- 01 Jan 1985
TL;DR: This study tested human performance on a real and virtual floor, as well as in a threedimensional (3D) virtual space, and modeled these results by a graph pyramid algorithm, which suggests that deterioration of performance in the 3D space can be attributed to geometrical relations between hierarchical clustering in a3D space and coarse-to-fine production of a tour.
1.8K
Restless bandits: activity allocation in a changing world
TL;DR: In this article, the Lagrange multiplier associated with this constraint defines an index which reduces to the Gittins index when projects not being operated are static, and arguments are advanced to support the conjecture that, for m and n large in constant ratio, the policy of operating the m projects of largest current index is nearly optimal.
1.3K
A compile-time scheduling heuristic for interconnection-constrained heterogeneous processor architectures
G.C. Sih,Edward A. Lee +1 more
TL;DR: The authors present a compile-time scheduling heuristic called dynamic level scheduling, which accounts for interprocessor communication overhead when mapping precedence-constrained, communicating tasks onto heterogeneous processor architectures with limited or possibly irregular interconnection structures.
951