An experimental evaluation of processor pool-based scheduling for shared-memory NUMA multiprocessors

Brecht, Tim

doi:10.1007/3-540-63574-2_20

Cited by 5 publications

(6 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used this application structure and the data collected from the DAS in the simulations, replacing the distribution D of job component sizes chosen before with the two job sizes considered above: (2,2,2,2) and (4,4,4,4) and assessed the performance of the system (response time as a function of the system utilization) for ordered requests. Because the job components are equal, identical performance results would be obtained for unordered requests.…”

Section: Simulation Resultsmentioning

confidence: 99%

See 1 more Smart Citation

The Influence of Communication on the Performance of Co-allocation

Bucur

Epema

2001

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

In systems consisting of multiple clusters of processors interconnected by relatively slow connections such as our Distributed ASCI 1 Supercomputer (DAS), jobs may request coallocation, i.e., the simultaneous allocation of processors in different clusters. The performance of co-allocation may be severely impacted by the slow intercluster connections, and by the types of job requests. We distinguish different job request types ranging from ordered requests that specify the numbers of processors needed in each of the clusters, to flexible requests that only specify a total. We simulate multicluster systems with the FCFS policy-and with two policies for placing a flexible request, one tries to balance cluster loads and one tries to fill clusters completely-to determine the response times under workloads consisting of a single or of different request types for different communication speeds across the intercluster connections. In addition to a synthetic workload, we also consider a workload derived from measurements of a real application on the DAS. We find that the communication speed difference has a severe impact on response times, that a relatively small amount of capacity is lost due to communication, and that for a mix of request types, the performance is determined not only by the separate behaviours of the different types of requests, but also by the way in which they interact.

show abstract

Section: Simulation Resultsmentioning

confidence: 99%

“…In [4], the influence of splitting the processors of multiprogrammed, shared-memory NUMA multiprocessors into groups on the performance of the system is studied. It is assumed that a job starts as a single process, and that it may grow by creating additional processes.…”

Section: Related Workmentioning

confidence: 99%

The Influence of Communication on the Performance of Co-allocation

Bucur

Epema

2001

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

show abstract

“…Actual evaluations of scientific applications with representative datasets often take prohibitively long time. This problem is exacerbated when evaluating longer workloads or conducting parameter space explorations [8]. In some cases, researchers prefer to use application kernels instead of actual applications, but care must be taken to differentiate those from the real applications [5].…”

Section: Pitfall 9 Eee Measuring Biased Unrepresentative or Overly mentioning

confidence: 99%

Pitfalls in Parallel Job Scheduling Evaluation

Frachtenberg

Feitelson

2005

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. There are many choices to make when evaluating the performance of a complex system. In the context of parallel job scheduling, one must decide what workload to use and what measurements to take. These decisions sometimes have subtle implications that are easy to overlook. In this paper we document numerous pitfalls one may fall into, with the hope of providing at least some help in avoiding them. Along the way, we also identify topics that could benefit from additional research.

show abstract

“…The influence of splitting the processors into groups on the performance of the system was also studied in [3]. A technique for operating system schedulers called processor pool-based scheduling, designed to assign the processes of parallel applications in multiprogrammed, shared-memory NUMA multiprocessors, is presented and evaluated.…”

Section: Related Workmentioning

confidence: 99%

“…The necessity of dividing the processors into pools in order to simplify the scheduling decisions is discussed in the literature [3]. In the case of multiclusters, the division is natural and is imposed by the architecture of the system, and not by the scheduler.…”

Section: Introductionmentioning

confidence: 99%

The Influence of the Structure and Sizes of Jobs on the Performance of Co-allocation

Bucur

Epema

2000

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

Over the last decade, much research in the area of scheduling has concentrated on single-cluster systems. Less attention has been paid to multicluster systems, although they are gaining more and more importance in practice. We propose a model for scheduling rigid jobs consisting of multiple components in multicluster systems by pure space sharing, based on the Distributed ASCI Supercomputer. Using simulations, we asses the influence of the structure and sizes of the jobs on the system's performance, measured in terms of the average response time and the maximum utilization. We consider three types of requests, total requests, unordered requests and ordered requests, and compare their effect on the system's performance for two scheduling policies, First Come First Served, and Fit Processors First Served, which allows the scheduler to look further in the queue for jobs that fit. These types of job requests are differentiated by the restrictions they impose on the scheduler and by the form of co-allocation used. The results show that the performance improves with decreasing average job size and when fewer restrictions are imposed on the scheduler.

show abstract

An experimental evaluation of processor pool-based scheduling for shared-memory NUMA multiprocessors

Cited by 5 publications

References 35 publications

The Influence of Communication on the Performance of Co-allocation

The Influence of Communication on the Performance of Co-allocation

Pitfalls in Parallel Job Scheduling Evaluation

The Influence of the Structure and Sizes of Jobs on the Performance of Co-allocation

Contact Info

Product

Resources

About