Adaptive Scheduling for Master-Worker Applications on the Computational Grid

Heymann, Elisa; Senar, Miquel A.; Luque, Emilio; Livny, Miron

doi:10.1007/3-540-44444-0_20

Cited by 77 publications

(51 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…BoT applications composed of independent tasks with file sharing have been studied in several papers during the last years [4,6,[8][9][10][11]15,[17][18][19]. Their relevance has motivated the development of specialized environments which aim to facilitate the execution of large BoT applications on computational grids and clusters, such as the AppLeS Parameter-Sweep Template (APST) [3] and MyGrid [7].…”

Section: Related Workmentioning

confidence: 99%

“…In this section we build on the work done previously by several research groups [4,6,[8][9][10][11]15,[17][18][19] in order to propose a novel oblivious algorithm, dubbed Dynamic Clustering, using the input file affinity (I aff ) measure. In particular, the scheduling algorithm presented in this section is an evolution of the algorithm described in [18].…”

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

“…Compute average execution time on processor P i ; (6) Update task queue; (7) Abort still running replicas of completed tasks; (8) If Processor P i is idle (9) If there are unfinished groups not yet replicated on slower processors (10) Replicate unfinished group into processor P i ; } Fig. 2.…”

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

“…It is also possible to obtain information about the current load of a machine by measuring the number of tasks still to be executed. In [10], authors also use information collected dynamically to schedule tasks of iterative parallel applications on processors. -(8) If a machine becomes idle, send a replica of the remaining tasks (not yet replicated) of the machine with the largest amount of unfinished computation to be executed on the idle machine.…”

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

See 3 more Smart Citations

Improving scalability of Bag-of-Tasks applications running on master–slave platforms

Silva

Senger

2009

Parallel Computing

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

Section: Using the I Aff Measure To Schedule Bot Applicationsmentioning

confidence: 99%

See 2 more Smart Citations

Improving scalability of Bag-of-Tasks applications running on master–slave platforms

Silva

Senger

2009

Parallel Computing

View full text Add to dashboard Cite

“…This architecture (also referred to as task farming architecture) consists of one master entity and multiple workers entities, wherein the master entity decomposes the problem into small tasks, distributes these tasks among a farm of worker processes and gathers the partial results to produce the final result of the computation; and the worker entities receive message from the master with the next task, process the task and send back the result to the master (Heymann et al, 2000). WinGrid implements this "push" approach (master pushes the job to the workers) by starting a server process for each worker.…”

Section: Wingrid Architecturementioning

confidence: 99%

Speeding up simulation applications using WinGrid

Mustafee

Taylor

2009

Concurrency and Computation

View full text Add to dashboard Cite

Abstract. The vision of grid computing is to make computational power, storage capacity, data and applications available to users as readily as electricity and other utilities. Grid infrastructures and applications have traditionally been geared towards dedicated, centralized, high performance clusters running on UNIX "flavour" operating systems (commonly referred to as cluster-based grid computing). This can be contrasted with desktopbased grid computing which refers to the aggregation of non-dedicated, de-centralized, commodity PCs connected through a network and running (mostly) the Microsoft Windows operating system. Large scale adoption of such Windows-based grid infrastructure may be facilitated via grid-enabling existing Windows applications. This paper presents the WinGrid approach to grid enabling existing Windows-based Commercial-Off-The-Shelf (COTS) simulation packages (CSPs). Through the use of two case studies developed in conjunction with a major automotive company and a leading investment bank respectively, the contribution of this paper is the demonstration of how experimentation with the CSP Witness (Lanner Group) and the CSP Analytics (SunGard Corporation) can achieve speedup when using WinGrid middleware on both dedicated and non-dedicated grid nodes. It is hoped that this research would facilitate wider acceptance of desktop grid computing among enterprises interested in a low-intervention technological solution to speeding up their existing simulations.Keywords. Grid computing, desktop grids, grid middleware, commercial-off-the-shelf simulation packages, discrete-event simulation, Monte Carlo simulation. IntroductionGrids are sharing environments implemented via the deployment of a persistent, standardsbased service infrastructure that supports the creation of distributed communities and sharing of resources like computers, storage space, sensors, software applications and data between them (Foster and Iamnitchi, 2003). These distributed communities, frequently referred to as virtual organizations, or virtual enterprises, comprise of a group of individuals and/or institutions engaged in some joint work who share resources based on strict sharing policies that define what is shared, who is allowed to share and the conditions under which such sharing occurs (Foster et al., 2002).Computer simulation has the potential to benefit from sharing access to computing resources, storage capacities and research equipments provided by grid computing. Examples of large scale grid-based simulation projects include the Earth Grid System (Bernholdt et al., 2005) and NEESgrid (Spencer et al., 2004). The creation of such applications typically requires the installation of complex supporting software (like Globus) and an in-depth knowledge of how this complex supporting software works (Jaesun et al., 2003).The exponential growth of global computer ownership, local networks and Internet connectivity, coupled with the fact that desktop PCs in corporate and home environments are heavily under utilized, has given r...

show abstract

Adaptive structured parallelism for distributed heterogeneous architectures: a methodological approach with pipelines and farms

González–Vélez

Cole

2010

Concurrency and Computation

View full text Add to dashboard Cite

SUMMARYAlgorithmic skeletons abstract commonly used patterns of parallel computation, communication, and interaction. Based on the algorithmic skeleton concept, structured parallelism provides a high-level parallel programming technique that allows the conceptual description of parallel programs while fostering platform independence and algorithm abstraction. This work presents a methodology to improve skeletal parallel programming in heterogeneous distributed systems by introducing adaptivity through resource awareness. As we hypothesise that a skeletal program should be able to adapt to the dynamic resource conditions over time using its structural forecasting information, we have developed adaptive structured parallelism (ASPARA). ASPARA is a generic methodology to incorporate structural information at compilation into a parallel program, which will help it to adapt at execution. ASPARA comprises four phases: programming, compilation, calibration, and execution. We illustrate the feasibility of this approach and its associated performance improvements using independent case studies based on two algorithmic skeletons-the task farm and the pipeline-evaluated in a non-dedicated heterogeneous multi-cluster system.

show abstract

Adaptive Scheduling for Master-Worker Applications on the Computational Grid

Cited by 77 publications

References 11 publications

Improving scalability of Bag-of-Tasks applications running on master–slave platforms

Improving scalability of Bag-of-Tasks applications running on master–slave platforms

Speeding up simulation applications using WinGrid

Adaptive structured parallelism for distributed heterogeneous architectures: a methodological approach with pipelines and farms

Contact Info

Product

Resources

About