Interactive Launch of 16,000 Microsoft Windows Instances on a Supercomputer

Jones, Michael; Kepner, Jeremy; Orchard, Bradley J.; Reuther, Albert; Arcand, William; Bestor, David; Bergeron, Bill; Byun, Chansup; Gadepally, Vijay; Houle, Michael E.; Hubbell, Matthew; Klein, Anna; Milechin, Lauren; Mullen, Julia; Prout, Andrew; Rosa, Antonio; Samsi, Siddharth; Yee, Charles; Michaleas, Peter

doi:10.1109/hpec.2018.8547782

Cited by 3 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Operating System Translators allow the execution of applications built for OSs different from the system host, e.g., WINE (Jones et al, 2018;Wine, 2022), WABI (Oracle, 2018), Lxrun (2022), Visual MainWin (Fisher et al, 2006), and Vcuda (Balis et al, 2021;S. Lin et al, 2009).…”

Section: • Multitasking Os Is Divided Into Operating Systemmentioning

confidence: 99%

A Survey of Virtualization Technologies: Towards a New Taxonomic Proposal

et al. 2022

View full text Add to dashboard Cite

At present, there is a proliferation of virtualization technologies (VTs), which are part of the basic and underlying infrastructure of popular cloud computing. Those interested in VTs are faced with a non-unified volume of information and various approaches to modes of operation, classification structures, and the performance implications of these technologies. This makes it difficult to decide which type of VT is appropriate for a particular context. Therefore, this paper reviews the state of the art on VT taxonomic models. Methodologically, a literature review is carried out to identify VT classification models, recognizing their features and weaknesses. With this in mind, a new taxonomy of virtualization technologies is proposed, which responds to the weaknesses identified in the analyzed schemes. The new VT taxonomy combines the Abstraction Level and Virtual Machine Type approaches, providing the reader with a means to visualize VTs. In doing so, the reader can locate the level of abstraction at which each VT is developed, in addition to the type of machine projected, whether it is a complete system or an execution environment for processes. The proposed taxonomy can be used in the academic environment to facilitate teaching processes or in the business environment to facilitate decision-making when implementing VTs.

show abstract

Section: • Multitasking Os Is Divided Into Operating Systemmentioning

confidence: 99%

A Survey of Virtualization Technologies: Towards a New Taxonomic Proposal

et al. 2022

View full text Add to dashboard Cite

show abstract

“…In fact, one of us reported that it is common for a large cloud to be managing 10 8 Spark tasks at a time [7]. Interactive parallel computing systems require launching thousands of jobs in a few seconds [21,32]. Obviously, one needs "gang" scheduling and straggler mitigation support.…”

Section: Introductionmentioning

confidence: 99%

Dbos

Skiadopoulos

Kraft

et al. 2021

Proc. VLDB Endow.

Self Cite

View full text Add to dashboard Cite

This paper lays out the rationale for building a completely new operating system (OS) stack. Rather than build on a single node OS together with separate cluster schedulers, distributed filesystems, and network managers, we argue that a distributed transactional DBMS should be the basis for a scalable cluster OS. We show herein that such a database OS (DBOS) can do scheduling, file management, and inter-process communication with competitive performance to existing systems. In addition, significantly better analytics can be provided as well as a dramatic reduction in code complexity through implementing OS services as standard database queries, while implementing low-latency transactions and high availability only once.

show abstract

Optimizing Xeon Phi for Interactive Data Analysis

Byun

Klein

Milechin³

et al. 2019

2019 IEEE High Performance Extreme Computing Conference (HPEC)

Self Cite

View full text Add to dashboard Cite

The Intel Xeon Phi manycore processor is designed to provide high performance matrix computations of the type often performed in data analysis. Common data analysis environments include Matlab, GNU Octave, Julia, Python, and R. Achieving optimal performance of matrix operations within data analysis environments requires tuning the Xeon Phi OpenMP settings, process pinning, and memory modes. This paper describes matrix multiplication performance results for Matlab and GNU Octave over a variety of combinations of process counts and OpenMP threads and Xeon Phi memory modes. These results indicate that using KMP_AFFINITY=granlarity=fine, taskset pinning, and all2all cache memory mode allows both Matlab and GNU Octave to achieve 66% of the practical peak performance for process counts ranging from 1 to 64 and OpenMP threads ranging from 1 to 64. These settings have resulted in generally improved performance across a range of applications and has enabled our Xeon Phi system to deliver significant results in a number of real-world applications.

show abstract

Interactive Launch of 16,000 Microsoft Windows Instances on a Supercomputer

Cited by 3 publications

References 27 publications

A Survey of Virtualization Technologies: Towards a New Taxonomic Proposal

A Survey of Virtualization Technologies: Towards a New Taxonomic Proposal

Dbos

Optimizing Xeon Phi for Interactive Data Analysis

Contact Info

Product

Resources

About