Entity

Time filter

Source Type

Amsterdam-Zuidoost, Netherlands

Dumitru C.,University of Amsterdam | Oprescu A.-M.,University of Amsterdam | Zivkovic M.,University of Amsterdam | Van Der Mei R.,Center for Mathematics and Informatics | And 2 more authors.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2014

Cloud hosting services offer computing resources which can scale along with the needs of users. When access to data is limited by the network capacity this scalability also becomes limited. To investigate the impact of this limitation we focus on bags-of-tasks where task data is stored outside the cloud and has to be transferred across the network before task execution can commence. The existing bags-of-tasks estimation tools are not able to provide accurate estimates in such a case. We introduce a queuing-network inspired model which successfully models the limited network resources. Based on the Mean-Value Analysis of this model we derive an efficient procedure that results in an estimate of the makespan and the executions costs for a given configuration of cloud virtual machines. We compare the calculated Pareto set with measurements performed in a number of experiments for real-world bags-of-tasks and validate the proposed model and the accuracy of the estimated configurations. © 2014 Springer International Publishing Switzerland.


Zwart S.P.,Leiden University | Bedorf J.,Center for Mathematics and Informatics
Computer | Year: 2015

Optimizations for individual N-body techniques allow the simulation of collisonal or collisionless systems, but not both together. Hybrid code running on GPUs meets this requirement, and enabled the efficient and accurate simulation of 11 interacting galaxies with a massive black hole in each of their nuclei. © 2015 IEEE.


Bankovic A.,University of Belgrade | Dujko S.,University of Belgrade | Dujko S.,Center for Mathematics and Informatics | Dujko S.,James Cook University | And 4 more authors.
Journal of Physics: Conference Series | Year: 2011

Transport properties of positron swarms drifting and diffusing in neutral gases under the influence of crossed electric and magnetic fields are investigated using a multi-term theory for solving the Boltzmann equation and Monte Carlo simulation technique. In the presence of magnetic fields the number of transport properties is increased compared to the situation when the positron swarm is acted on solely by an electric field. Since the longitudinal and transverse components of the drift velocity show different sensitivities with respect to the strength of the magnetic field, it is found that the negative differential conductivity effect in a crossed field configuration can be controlled through the variation of the magnetic field strengths. Various diffusion tensor elements also exhibit different sensitivities with respect to the magnetic field and also with respect to the positronium (Ps) formation process.


Petrovic Z.Lj.,University of Belgrade | Bankovic A.,University of Belgrade | Marjanovic S.,University of Belgrade | Suvakov M.,University of Belgrade | And 6 more authors.
Journal of Physics: Conference Series | Year: 2011

In this paper we give a review of two recent developments in positron transport, calculation of transport coefficients for a relatively complete set of collision cross sections for water vapour and for application of they Monte Carlo technique to model gas filled subexcitation positron traps such as Penning Malmberg Surko (Surko) trap. Calculated transport coefficients, very much like those for argon and other molecular gases show several new kinetic phenomena. The most important is the negative differential conductivity (NDC) for the bulk drift velocity when the flux drift velocity shows no sign of NDC. These results in water vapour are similar to the results in argon or hydrogen. The same technique that has been used for positron (and previously electron) transport may be applied to model development of particles in a Surko trap. We have provided calculation of the ensemble of positrons in the trap from an initial beam like distribution to the fully thermalised distribution. This model, however, does not include plasma effects (interaction between charged particles) and may be applied for lower positron densities. © 2011 IOP Publishing Ltd.


Li C.,Center for Mathematics and Informatics | Ebert U.,Center for Mathematics and Informatics | Ebert U.,TU Eindhoven | Hundsdorfer W.,Center for Mathematics and Informatics | Hundsdorfer W.,Radboud University Nijmegen
Journal of Computational Physics | Year: 2010

Streamers are the first stage of sparks and lightning; they grow due to a strongly enhanced electric field at their tips; this field is created by a thin curved space charge layer. These multiple scales are already challenging when the electrons are approximated by densities. However, electron density fluctuations in the leading edge of the front and non-thermal stretched tails of the electron energy distribution (as a cause of X-ray emissions) require a particle model to follow the electron motion. But present computers cannot deal with all electrons in a fully developed streamer. Therefore, super-particle have to be introduced, which leads to wrong statistics and numerical artifacts. The method of choice is a hybrid computation in space where individual electrons are followed in the region of high electric field and low density while the bulk of the electrons is approximated by densities (or fluids). We here develop the hybrid coupling for planar fronts. First, to obtain a consistent flux at the interface between particle and fluid model in the hybrid computation, the widely used classical fluid model is replaced by an extended fluid model. Then the coupling algorithm and the numerical implementation of the spatially hybrid model are presented in detail, in particular, the position of the model interface and the construction of the buffer region. The method carries generic features of pulled fronts that can be applied to similar problems like large deviations in the leading edge of population fronts, etc. © 2009 Elsevier Inc. All rights reserved.

Discover hidden collaborations