Towards an efficient single system image cluster operating system

Morin, Christine; Gallard, Pascal; Lottiaux, Renaud; Vallée, Geoffroy

doi:10.1016/s0167-739x(03)00170-5

Cited by 30 publications

(20 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kerrighed [68,69] is a single-system image implementation for Linux, developed at INRIA during the early-to-mid 2000s. It is implemented as a small (less than 200 lines) patch to the Linux kernel itself along with a number of kernel modules and supporting userland tools.…”

Section: Kerrighedmentioning

confidence: 99%

Single system image: A survey

Healy

Lynn

Barrett

et al. 2016

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

Section: Kerrighedmentioning

confidence: 99%

Single system image: A survey

Healy

Lynn

Barrett

et al. 2016

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

“…Note that when child processes are created, at first, they all run on the access point node (since that is where the father process is issuing the fork()-operation). It is the Kerrighed scheduler, 22 which then uses preemptive process migration to distribute the processes appropriately to other nodes. On the other hand, the Debian Linux kernel's completely fair scheduler is responsible for the distribution of processes on a single node over multiple local CPU cores.…”

Section: Frequency Per Evaluation Of Equation 12 Together With Equatimentioning

confidence: 99%

“…21 Atlas takes great advantage of this and realizes Gemm as a portable high-performance routine adapting to the machines that run it, depending on the system, often getting close to theoretical peak performance. 20 In order to further enhance the computational efficiency of the multiple evaluation of Equation 12 together with Equation 5, the authors resorted to the implementation of a Kerrighed single system image 22 (SSI) computing system. An SSI cluster is a clustered computer system that, to the user, behaves as a single high-performance multiprocessor computer while effectively being a cluster of interconnected single-or multiprocessor computers sharing their resources transparently to the operator.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Fracture safety of double-porous hydroxyapatite biomaterials

DejacoAlexander

Komlev

Jaroszewicz

et al. 2016

Bioinspired, Biomimetic and Nanobiomaterials

View full text Add to dashboard Cite

Biological requirements call for substantial porosities in clinical biomaterials -challenging the mechanical integrity and strength of the latter. In this study, the authors resort to quantitative engineering principles to assess the fracture safety of double-porous hydroxyapatite ceramics: micro-computed tomography scans give access to the morphology of macropores at the submillimeter scale, as well as to voxel-specific microporosities. Advanced micromechanics of porous ceramics with needle-shaped elementary units then allows for translating voxel-specific microporosities to corresponding elasticity and strength properties, as well as to macro-to-micro scale transition ('concentration') tensors.These mechanical properties and tensors are fed into a large-scale finite-element model of a biomaterial granule as used for mandibular tissue regeneration. Loading the granule in splitting mode, up to physiological strain, evidences stress concentrations at the loaded poles and close to internal macropores and cracks. A parallel computing-supported subvoxel analysis of needle orientations evidences that in highly loaded regions, the intravoxel 'single crystals' oriented perpendicular to the loading direction undergo the most unfavorable loading. Still, only 0·6% of the finiteelements show stresses indicating failure, and the mean safety factor against fracture is as high as 7. This analysis confirms, from an engineering science viewpoint, the successful use of the investigated biomaterials in clinical practice. (j, ϑ) second-order microscopic stress tensor of a needle-shaped hydroxyapatite phase orientated in the (j, ϑ) direction s HA,NN (j, ϑ) normal component of stress tensor r HA (j, ϑ) in needle direction (j, ϑ) s HA,Nn (j, ϑ, y) shear component of stress tensor r HA (j, ϑ) on planes orthogonal to the needle direction (j, ϑ) s crit HA Notation

show abstract

“…Condor [20] is one example for such a system implemented in user space, MOSIX [3] is a module for the Linux kernel which allows migrations of single processes. Kerrighed [21] goes beyond this and provides a single system image of a workstation cluster. It supports migration of processes and also communication via shared memory.…”

Section: Introductionmentioning

confidence: 99%

Load Balancing in the Bulk-Synchronous-Parallel Setting using Process Migrations

Bonorden

2007

2007 IEEE International Parallel and Distributed Processing Symposium

View full text Add to dashboard Cite

The Paderborn University BSP (PUB) library is a powerful C library that supports the development of bulk synchronous parallel programs for various parallel machines. To utilize idle times on workstations for parallel computations, we implement virtual processors using processes. These processes can be migrated to other hosts, when the load of the machines changes. In this paper we describe the implementation for a Linux workstation cluster. We focus on process migration and show first benchmarking results.

show abstract

Towards an efficient single system image cluster operating system

Cited by 30 publications

References 20 publications

Single system image: A survey

Single system image: A survey

Fracture safety of double-porous hydroxyapatite biomaterials

Load Balancing in the Bulk-Synchronous-Parallel Setting using Process Migrations

Contact Info

Product

Resources

About