Open Job Management Architecture for the Blue Gene/L Supercomputer

Aridor, Yariv; Domany, Tamar; Goldshmidt, O.; Kliteynik, Yevgeny; Moreira, José E.; Shmueli, Edi

doi:10.1007/11605300_4

Cited by 12 publications

(8 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Examples include high bandwidth interconnection switches (e.g. SP2 switch [12], InfiniBand [10]) hardware multi-threading, and Blue Gene [6,5]. Supported software features include AIX's WLM [1], and various Linux distributions [4].…”

Section: Loadleveler Overviewmentioning

confidence: 99%

Scalability Analysis of Job Scheduling Using Virtual Nodes

Bobroff

Coppinger²,

Fong

et al. 2009

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

Abstract. It is important to identify scalability constraints in existing job scheduling software as they are applied to next generation parallel systems. In this paper, we analyze the scalability of job scheduling and job dispatching functions in the IBM LoadLeveler job scheduler. To enable this scalability study, we propose and implement a new virtualization method to deploy different size LoadLeveler clusters with minimal number of physical machines. Our scalability studies with the virtualization show that the LoadLeveler resource manager can comfortably handle over 12,000 compute nodes, the largest scale we have tested so far. However, our study shows that the static resource matching in the scheduling cycle and job object processing during the hierarchical job launching are two impediments for the scalability of LoadLeveler.

show abstract

Section: Loadleveler Overviewmentioning

confidence: 99%

Scalability Analysis of Job Scheduling Using Virtual Nodes

Bobroff

Coppinger²,

Fong

et al. 2009

Job Scheduling Strategies for Parallel Processing

View full text Add to dashboard Cite

show abstract

“…Little has changed in the external appearance of mpirun from BG/L. As described in [1], it acts as a proxy, or shadow of the job executing on the BG/P core. This means when input is sent to mpirun by the user, it is routed to the job executing on the BG/P core appropriately.…”

Section: Mpirunmentioning

confidence: 99%

“…The authors of [1] describe the job management architecture of BG/L in detail, concluding that it is capable of supporting job management systems developed internally at IBM such as LoadLeveler, as well as third party systems such as SLURM [4]. They define three openness characteristics.…”

Section: Related Workmentioning

confidence: 99%

A Flexible Resource Management Architecture for the Blue Gene/P Supercomputer

Miller

Megerian

Allen

et al. 2007

2007 IEEE International Parallel and Distributed Processing Symposium

View full text Add to dashboard Cite

Blue Gene R /P is a massively parallel supercomputer intended as the successor to Blue Gene/L. It leverages much of the existing architecture of its predecessor to provide scalability up to a petaflop of peak computing power. The resource management software for such a large parallel system faces several challenges, including system fragmentation due to partitioning, presenting resource usage information using a polling or event driven model, and acting as a barrier between external resource management systems and the Blue Gene/P core. This paper describes how the Blue Gene/P resource management architecture is extremely flexible by providing multiple methodologies for obtaining resource usage information to make scheduling decisions. Three distinctly separate resource management services will be described. First, the Bridge API, a full-featured API suitable for fine tuning scheduling and allocation decisions. Second, a light-weight Allocator API for allocating resources without substantial development costs. And lastly, configuring the system into static partitions. Job scheduling strategies utilizing each of the methods will be discussed.

show abstract

“…The mesh has been one of the most common networks for multicomputers due to its simplicity, scalability, structural regularity, and ease of implementation [2,5]. It has been used as the underlying network in a number of practical and experimental parallel machines, such as iWARP [3], the IBM BlueGene/L [18], and Delta Touchstone [6].…”

Section: Introductionmentioning

confidence: 99%