The workload on parallel supercomputers: modeling the characteristics of rigid jobs

By Uri Lublin, Dror G. Feitelson. J. Parallel Distributed Comput., 2003

488 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

An Analysis of Traces from a Production MapReduce Cluster (2010, result): “Our analysis of MapReduce workloads corroborates observations from previous studies of workloads on parallel computer systems [18], [23], [24].”
Evaluating the cost-benefit of using cloud computing to extend the capacity of clusters (2009, methodology): “For this experiment, the site’s workloads have been generated using Lublin and Feitelson’s model [22], here referred to as Lublin99.”
An Early Performance Analysis of Cloud Computing Services for Scientific Computing (2008, unknown): “Also, scientific computing is traditionally a high-utilization workload, with production grids often running at over 80% utilization [17] and parallel production infrastructures (PPIs) averaging over 60% utilization [ 26 ].”
Workload analysis of a large-scale key-value store (2012, methodology): “Of those, a good survey of the methods is presented in Lublin’s and Feitelson’s analysis of supercomputer workloads [17, 24].”
The Grid Workloads Archive (2008, methodology): “Several workload models ar upported, including the Lublin-Feitelson model [14].”
Deconstructing Amazon EC2 Spot Instance Pricing (2011, methodology): “[2011] resorted to a grid and parallel systems model [Lublin and Feitelson 2003] with adapted runtime parameters to describe cloud workloads.”