Browse ORBi

- What it is and what it isn't
- Green Road / Gold Road?
- Ready to Publish. Now What?
- How can I support the OA movement?
- Where can I learn more?

ORBi

On the performance of an overlapping-domain parallelization strategy for Eulerian-Lagrangian Multiphysics software Pozzetti, Gabriele ; Besseron, Xavier ; Rousset, Alban et al in AIP Conference Proceedings (in press) In this work, a strategy for the parallelization of a two-way CFD-DEM coupling is investigated. It consists on adopting balanced overlapping partitions for the CFD and the DEM domains, that aims to reduce ... [more ▼] In this work, a strategy for the parallelization of a two-way CFD-DEM coupling is investigated. It consists on adopting balanced overlapping partitions for the CFD and the DEM domains, that aims to reduce the memory consumption and inter-process communication between CFD and DEM. Two benchmarks are proposed to assess the consistency and scalability of this approach, coupled execution on 252 cores shows that less than 1\% of time is used to perform inter-physics data exchange. [less ▲] Detailed reference viewed: 7 (1 UL)Multiscale model of sintering: diffusion and plastic flow Kabore, Brice Wendlassida ; Peters, Bernhard Scientific Conference (2017, September 27) Impacting particles or static aggregated particles at high temperature may undergo a permanent change of shape modifying the microstructure. Two particles in contact can develop some bonds within sub ... [more ▼] Impacting particles or static aggregated particles at high temperature may undergo a permanent change of shape modifying the microstructure. Two particles in contact can develop some bonds within sub-second time. This fast sintering force in the particular case of the snow contribute to the rheological behavior and grain rearrangement [1]. Understanding the kinetics of sintering in granular material is of great importance in some engineering applications. For decades, diffusional processes have received more attention in investigations related to the mechanisms behind sintering [2]. Some works have suggested that the plastic flow might be neglected in sintering process for stresses are not high enough to cause dislocation. However, some studies have showed that stresses experienced in fine particles necks can be high enough and even lead to plasticity driven sintering. The importance of each mechanism in the sintering process may lie in the temporal and spatial scale of interest. Increasing importance is being accorded to the role of plastic flow in sintering. however, several investigations have proved that the conventional plasticity theory may fail to predict plastic activity at micro-scale, The objective of this work is to develop adequate computational model that includes instantaneous and time-dependent plastic flow at micro-scale. We aim at extending existing models of sintering and plasticity to cope with multiple spatial and temporal scales simulations using Extended Discrete Element Method. The numerical results are compare to experimental data on snow. [less ▲] Detailed reference viewed: 21 (2 UL)Comparing Broad-Phase Interaction Detection Algorithms for Multiphysics DEM Applications Rousset, Alban ; Mainassara Chekaraou, Abdoul Wahid ; Liao, Yu-Chung et al Scientific Conference (2017, September) Collision detection is an ongoing source of research and optimization in many fields including video-games and numerical simulations [6, 7, 8]. The goal of collision detection is to report a geometric ... [more ▼] Collision detection is an ongoing source of research and optimization in many fields including video-games and numerical simulations [6, 7, 8]. The goal of collision detection is to report a geometric contact when it is about to occur or has actually occurred. Unfortunately, detailed and exact collision detection for large amounts of objects represent an immense amount of computations, naivly n 2 operation with n being the number of objects [9]. To avoid and reduce these expensive computations, the collision detection is decomposed in two phases as it shown on Figure 1: the Broad-Phase and the Narrow-Phase. In this paper, we focus on Broad-Phase algorithm in a large dynamic three-dimensional environment. We studied two kinds of Broad-Phase algorithms: spatial partitioning and spatial sorting. Spatial partitioning techniques op- erate by dividing space into a number of regions that can be quickly tested against each object. Two types of spatial partitioning will be considered: grids and trees. The grid-based algorithms consist of a spatial partitioning processing by dividing space into regions and testing if objects overlap the same region of space. And this reduces the number of pairwise to test. The tree-based algorithms use a tree structure where each node spans a particular space area. This reduces the pairwise checking cost because only tree leaves are checked. The spatial sorting based algorithm consists of a sorted spatial ordering of objects. Axis-Aligned Bounding Boxes (AABBs) are projected onto x, y and z axes and put into sorted lists. By sorting projection onto axes, two objects collide if and only if they collide on the three axes. This axis sorting reduces the number of pairwise to tested by reducing the number of tests to perform to only pairs which collide on at least one axis. For this study, ten different Broad-Phase collision detection algorithms or framework have been considered. The Bullet [6], CGAL [10, 11] frameworks have been used. Concerning the implemented algorithms most of them come from papers or given implementation [less ▲] Detailed reference viewed: 21 (1 UL)Flow characteristics of metallic powder grains for additive manufacturing Peters, Bernhard ; Pozzetti, Gabriele in EPJ Web of Conferences (2017), 13001 Directed energy deposition technologies for additive manufacturing such as laser selective melting (SLM) or electron beam melting (EBM) is a fast growing technique mainly due to its flexibility in product ... [more ▼] Directed energy deposition technologies for additive manufacturing such as laser selective melting (SLM) or electron beam melting (EBM) is a fast growing technique mainly due to its flexibility in product de- sign. However, the process is a complex interaction of multi-physics on multiple length scales that are still not entirely understood. A particular challenging task are the flow characteristics of metallic powder ejected as jets from a nozzle and shielded by an inert turbulent gas flow. Therefore, the objective is to describe numerically the complex interaction between turbulent flow and powder grains. In order to include both several physical processes and length scales an Euler-Lagrange technology is applied. Within this framework powder is treated by the Discrete-Element-Method, while gas flow is described by Euler approaches as found in classical Compu- tational Fluid Dynamics (CFD). The described method succeeded in delivering more accuracy and consistency than a standard approach based on the volume averaging technique and therefore, is suited for the solution of problems within an engineering framework. [less ▲] Detailed reference viewed: 41 (13 UL)Investigating Multiphase flow Behavior in Trickle Bed Reactors using eXtended Discrete Element Method (XDEM) Baniasadi, Maryam ; Peters, Bernhard Scientific Conference (2017, May 08) The existence of multiphase flows through packed bed of solid particles in broad spectrum of engineering disciplines such as chemical industries, petroleum engineering, wastewater treatment is undeniable ... [more ▼] The existence of multiphase flows through packed bed of solid particles in broad spectrum of engineering disciplines such as chemical industries, petroleum engineering, wastewater treatment is undeniable. One frequently used reactor of this type is a trickle bed reactor that usually contains particulate phase of which the interstitial space is filled with gas and liquid phases. Based on the direction of the fluid flow they can be classified as cocurrent downflow trickle bed reactors, counter-current trickle bed reactors and cocurrent upflow packed bubble reactors. In these kind of problems numerical simulations can help to gain a better process understanding. In the current distribution, a numerical method so called Extended Discrete Element Method (XDEM) was applied to model multiphase flow through packed bed of solid particles which has the ability to be coupled to Computational Fluid Dynamics (CFD) through interphase momentum transfer. In this coupled solver the fluid phases are treated by CFD while the position and orientation of the particles in each CFD cell and the porosity distribution through packed bed are provided by XDEM. In order to validate the code, two important hydrodynamic parameters such as pressure drop and liquid hold up were investigated and satisfactory agreement between predicted and experimental data was achieved. The model results demonstrate enormous effect of solid particles on the deviation of fluid phases while passing through packed bed by investigating parameters such as velocity and drag force. [less ▲] Detailed reference viewed: 12 (1 UL)On the choice of a phase interchange strategy for a multiscale DEM-VOF Method Pozzetti, Gabriele ; Peters, Bernhard in AIP Conference Proceedings (2017), 1863 In this work a novel Multiscale DEM-VOF method is adopted to study three phase flows. It consists in solving the fluid momentum, mass conservation and the phase advection at a different scale with respect ... [more ▼] In this work a novel Multiscale DEM-VOF method is adopted to study three phase flows. It consists in solving the fluid momentum, mass conservation and the phase advection at a different scale with respect to the fluid-particle coupling problem. This allows the VOF scheme to resolve smaller fluid structures than a classic DEM-VOF method, and opens the possibility of adopting different volume interchange techniques. Two different volume interchange techniques are here described and compared with reference to high and low particle concentration scenarios. Considerations about the respective computational costs are also proposed. [less ▲] Detailed reference viewed: 141 (34 UL)Evaluating Erosion Patterns in an abrasive water jet cutting nozzle using XDEM Pozzetti, Gabriele ; Peters, Bernhard in Advances in Powder Metallurgy & Particulate Materials (2017) The objective of this work is to analyze particle-induced erosion within a nozzle for abrasive water jet cutting, through a Euler-Lagrange approach. In an abrasive water jet cutting device a high-speed ... [more ▼] The objective of this work is to analyze particle-induced erosion within a nozzle for abrasive water jet cutting, through a Euler-Lagrange approach. In an abrasive water jet cutting device a high-speed water jet is used to accelerate abrasive particles forming a turbulent mixture of water, entrained air and abrasive powders traveling at hundreds of meters per second. The focusing tube represents a key component, whose primary scope is to focus and stabilize the flow forming in the mixing chamber, in order to ensure optimal cutting performances of the device. Nevertheless, this nozzle often happens to be the first target of the erosive action of the flow. This phenomenon significantly shortens the operational life of a nozzle. The numerical approach proposed in this work aims to provide an insight to this very fast and disruptive phenomena that are difficult and expensive to be captured by purely experimental studies. [less ▲] Detailed reference viewed: 53 (14 UL)Multiscale-multiphysics approaches for engineering applications Pozzetti, Gabriele ; Peters, Bernhard in AIP Conference Proceedings (2017), 1863(1), 180001 Detailed reference viewed: 9 (0 UL)Application of the extended discrete element method (XDEM) in the melting of a single particle Baniasadi, Mehdi ; Baniasadi, Maryam ; Peters, Bernhard in Baniasadi, Mehdi (Ed.) Application of the extended discrete element method (XDEM) in the melting of a single particle (2016, July 19) In this contribution, a new method referred to as Extended Discrete Element Method (XDEM) is usedto model melting of a single particle in the fluid media. The XDEM as a Lagrangian-Eulerian framework is ... [more ▼] In this contribution, a new method referred to as Extended Discrete Element Method (XDEM) is usedto model melting of a single particle in the fluid media. The XDEM as a Lagrangian-Eulerian framework is the extension of Discrete Element Method (DEM) by considering thermodynamic state such as temperature distribution and is able to link with Computational Fluid Dynamics (CFD) for fluid phase. In order to provide more accurate results, multiscale method was used. The model is validated by comparing predicted results with existing experimental data for melting of a single ice particle in a water bath. In addition, the model has the capability to be extended to the packed bed of particles with different size and properties to produce different liquid phases. [less ▲] Detailed reference viewed: 15 (7 UL)ON THE INFLUENCE OF DIFFERENT MAPPING TECHNIQUES FOR A MULTISCALE APPROACH TO TURBULENT THREE-PHASE FLOWS Pozzetti, Gabriele ; Peters, Bernhard Scientific Conference (2016, June 06) In this work we investigate a multiscale approach for high Stokes number, turbulent three phase flows. It is widely proven that a straightforward application of Galerkin's method to problems characterized ... [more ▼] In this work we investigate a multiscale approach for high Stokes number, turbulent three phase flows. It is widely proven that a straightforward application of Galerkin's method to problems characterized by multiscale phenomena does not generally lead to robust numerical solutions. In this optic, multiscale methods are commonly adopted in order to provide solutions for complex problems in an highly efficient way. In certain problems it is convenient to identify multiple scales (more than 2), each characterized by its own characteristic spatial and temporal length. For this kind of problems a possible approach consists in completely resolving the coarse scales, partially resolving the middle scales, while analytically modeling the smallest. In turbulent three phase flows with high Stokes number, those can be identified respectively as the particle characteristic scale, the interface dynamic scale, and the turbulent fine scale. The coarse scale is here resolved through an Eulerian-Lagrangian approach that enables us to track the particle motion in a Lagrangian way. We partially resolve the middle-scale through the usage of a supporting domain where semi local variables are resolved. The solution of the middle-scale is based on the Volume of Fluid (VOF) technique in order to capture the dynamic interface, while turbulent phenomena are solved with a Large Eddy Simulation (LES) approach. The coarse-scale domain and the middle-scale domain must exchange informations and this process is obtained by mapping variables between the two fields. We will here show how the choice of the mapping technique largely affect the solution in therms of both accuracy and efficiency. A thoughtful study about the optimal mapping strategy could therefore be extremely beneficial in order to discover the most suitable scale-linking technique. The aim of this work is to investigate the effect of the adopted mapping technique on the resolved scale. Simulations with different Reynolds and Stokes number are proposed and compared, and conclusions about the consistency of the mapping technique are drawn. [less ▲] Detailed reference viewed: 70 (6 UL)CFD-XDEM FOR PREDICTING MULTIPHASE FLOW BEHAVIOR THROUGH POROUS MEDIA Baniasadi, Maryam ; Peters, Bernhard Scientific Conference (2016, June 05) An accurate description of fluid flow through porous media is very important to predict, design and optimize many industrial phenomena principally in condition where experimental studies are difficult to ... [more ▼] An accurate description of fluid flow through porous media is very important to predict, design and optimize many industrial phenomena principally in condition where experimental studies are difficult to perform. In these kind of problems numerical simulations can help to gain a better process understanding. During last decades many numerical approaches mainly Finite Volume Method (FVM) were applied to model different multiphase flows containing gas, liquid and solid phases. The solid phase may treat by continuous or discrete frameworks. In the former method which is based on Eulerian framework the solid phase is considered as a continuous phase like other fluid phases while in the second method which is based on Lagrangian framework, the solid phase is considered as separate particles. In this study, the flow behavior of several incompressible isotherm phases through solid particles was modelled. The model describes the motion of fluid flows such as gas and liquid phases using continuum approach by applying Computational Fluid Dynamics (CFD) as a numerical method and the solid particles by Lagrangian framework using so-called eXtended Discrete Element Method (XDEM). XDEM is a numerical simulation framework based on classical Discrete Element Methods (DEM) extended by consideration of thermophysical states. A combination of the two numerical methods was performed through momentum and mass exchange between fluid and solid phases which is called combined continuum discrete approach. The solid phase is considered as packed solid particles. The model results demonstrate enormous effect of solid particles on deviation of fluid phases while passing through particles by testing different drag force models. This model was applied to the dripping zone of blast furnace where the liquid phases of liquid iron and slag flow downward through coke particles and gas phase ascends upwards through the shaft which is classified as a counter-current multiphase packed bed reactor. The main goal of this project is to provide a solver which is able to treat several fluid phases through porous media using combined Eulerian-Lagrangian framework by exchanging data between this two approaches. [less ▲] Detailed reference viewed: 8 (1 UL)A DEM-LES-VOF METHOD FOR TURBULENT THREE PHASE FLOWS Pozzetti, Gabriele ; Peters, Bernhard Scientific Conference (2016, May 26) In this work a robust Computational Fluid Dynamic (CFD) - Discrete Element Method(DEM) coupling that can predict free-surface, turbulent flows is presented. A correct prediction of multiphase turbulent ... [more ▼] In this work a robust Computational Fluid Dynamic (CFD) - Discrete Element Method(DEM) coupling that can predict free-surface, turbulent flows is presented. A correct prediction of multiphase turbulent flows should ideally be able to capture the discrete dynamics of a dispersed phase (solid particles), and at the same time to take into account the evolution of possible fluid-dynamic instabilities. In this optic a CFD-DEM approach seems promising as it is able to combine the well developed CFD techniques for the study of free-surface flows with the accuracy of the Discrete Particle Method(DPM). A key point of the CFD-DEM method is the coupling between the discrete and the continuous phases. In particular the volume replacement between phases, and the interaction between the discrete phase and the continuous interface must be taken into account in order to perform accurate three phase simulations. In this work two different approaches to simulate the volume replacement between phases are presented and compared within a four way coupling with a Large Eddy Simulation(LES)-Volume Of Fluid(VOF) solver. The four-way coupled equations for the solid and the fluids are then presented, and some test cases provided in order to evaluate the accuracy of the new solver. Particular emphasis is posed to study the effects of the coupling on the interface dynamics and stability. The continuous two-phase solver used for the simulations is based on the OpenFoam® architecture, while the discrete phase solver is built on the XDEM code. [less ▲] Detailed reference viewed: 203 (8 UL)A PRELIMINARY STUDY ON THE STABILITY OF PARTICLE LADEN JETS THROUGH A FULLY COUPLED CFD-DEM SOLVER Peters, Bernhard ; Pozzetti, Gabriele Poster (2016, May 22) Jets are widely used in engineering applications. In material machinery, hydro-transportation systems as well as in chemical industry it is common to deal with a dispersed solid phase interacting with the ... [more ▼] Jets are widely used in engineering applications. In material machinery, hydro-transportation systems as well as in chemical industry it is common to deal with a dispersed solid phase interacting with the jet, and therefore creating a so-called slurry-jet or particle-laden jet. The stability of a jet is a key issue for many of these processes, still the underlying physics of this turbulent multiphase flow is highly complicated. Conventional CFD approaches have been proven satisfying for the study of the stability of two-phase jets. When a solid dispersed phase is present in the system, the stability problem gets more complicated and dependent on the solid phase dynamic. A possible solution for the problem is to extend the CFD solver capability through a correct coupling with a DEM solver. In this work a preliminary investigation on the potentialities of this kind of approach is presented and compared with a pure CFD approach. In particular the effect of the presence of differently sized particles in the jet is outlined and the influence of particle properties and concentration is investigated. Finally some considerations about the computational cost of different methods are proposed. The fluid phases are solved through an Eulerian finite volume (FV) multiphase solver based on the OpenFoam® libraries, and coupled with the XDEM code in order to treat the dispersed phase in a Lagrangian way. [less ▲] Detailed reference viewed: 105 (15 UL)A Combined Experimental and Numerical Approach to a Discrete Description of Indirect Reduction of Iron Oxide Peters, Bernhard ; ; et al in LA METALLURGIA ITALIANA (2016), (3), 49-54 Blast furnaces are complex counter-current reactors designed to reduce chemically iron oxides and melt them to liquid iron. The complex processes in blast furnace iron making involve various aspects of ... [more ▼] Blast furnaces are complex counter-current reactors designed to reduce chemically iron oxides and melt them to liquid iron. The complex processes in blast furnace iron making involve various aspects of thermodynamics, fluid dynamics, chemistry and physics. Physical, thermal and chemical phenomena occurring within the process are highly coupled in time and space. In order to generate a more detailed understanding of the indirect reduction of iron ore, the innovative approach of the Extended Discrete Element Method (XDEM) is applied. It describes the ore particle as discrete entities for which the thermodynamic state e.g. temperature and reduction degree through a reaction mechanism is described individually for each particle. The flow within the void space between the particles is represented by classical computational fluid dynamics that solves for the flow and temperature distribution including the composition of the gas phase. Ore particles and gas phase are tightly coupled by heat and mass transfer, that allows particles to heat up and to be provided with the reducing agent i.e. carbon monoxide. Reduction of iron oxide is predicted by a set of equilibrium reactions that represent the phase diagram of iron oxides at different oxidation levels. The reaction mechanism was validated by experimental data for a single ore particle for different temperatures. A comparison between measurements and predictions yielded good agreement so that reduction of iron oxide to iron was represented by a single mechanism including all reduction steps. The validated reaction mechanism was then applied to each particle of a packed bed that was exposed to define gas flow with its temperature and composition. The predicted results were also compared to experimental data and very good agreement was achieved. Due to the resolution of iron reduction on a particle level, detailed results of the entire reactor were obtained unveiling the underlying physics of the entire process. Results showed the reduction state of each particle during the entire period and additionally revealed the inhibiting influence of a non-uniform flow distribution. It provided regions of the packed bed with insufficient amounts of the reducing agent and thus, allowed identifying drawbacks for design and operation. [less ▲] Detailed reference viewed: 106 (10 UL)HPC or the Cloud: a cost study over an XDEM Simulation ; Besseron, Xavier ; Varrette, Sébastien et al in Proc. of the 7th International Supercomputing Conference in Mexico (ISUM 2016) (2016) Detailed reference viewed: 63 (8 UL)Berechnung des Transportes von Treibgut bei Hochwasser Peters, Bernhard ; Pozzetti, Gabriele ; Liao, Yu-Chung in 39. DRESDNER WASSERBAUKOLLOQUIUM (2016) Hochwasser hervorgerufen durch natürliche Ursachen wie Schneeschmelze oder durch bauliche Maßnahmen wie Flussbegradigung verursacht häufig eine Flutkatastrophe mit verheerenden Überschwemmungen. Zu den ... [more ▼] Hochwasser hervorgerufen durch natürliche Ursachen wie Schneeschmelze oder durch bauliche Maßnahmen wie Flussbegradigung verursacht häufig eine Flutkatastrophe mit verheerenden Überschwemmungen. Zu den schon katastrophalen Folgen von Hochwasser addieren sich häufig noch die Schäden von gefährliche Treibgut, das mit den Fluten mitgerissen wird und unter Umständen über weite Strecken transportiert wird. Mitgerissenes Treibgut kann zur Verklausung von Brücken führen oder auch Bauwerke zerstören. Um die Folgen eines Hochwassers einschließlich Transport von Treibgut abschätzen zu können, sind numerische Werkzeuge eine adäquate Ergänzung zu experimentellen Methoden, die oft mit einem sehr hohen Aufwand verbunden sind. Deshalb wird im vorliegenden Beitrag eine neue und innovative numerischer Ansatz vorgestellt, der den Transport von Treibgut bei Hochwasser aber auch bei Normalwasser beschreibt. Dazu werden die beiden numerischen Methoden beruhend auf einem diskreten und kontinuierlichem Ansatz gekoppelt. Letzterer beinhaltet die Euler Methoden, mit denen die Strömung des Wassers im Rahmen von bewährten Rechenmethoden der Computational Fluid Dynamik (CFD) bestimmt wird. Treibgut wird diskret betrachtet, in dem es mit der Diskreten Element Methode (DEM) beschreiben wird. Damit kann sowohl jedes einzelne Element des Treibgutes berücksichtigt werden als auch seine Eigenschaften wie Größe, Form und Gewicht. Innerhalb dieses Ansatzes können die Kontaktkräfte zwischen den einzelnen Elementen des Treibgutes berechnet werden, mit denen sich Geschwindigkeit, Position und Orientierung des Treibgutes bestimmen lassen. Zusätzlich wird über eine Kopplung zur fluiden Phase der Einfluss sowohl der Wassergeschwindigkeit als des Auftriebs mit berücksichtigt. [less ▲] Detailed reference viewed: 108 (16 UL)A discrete-continuous approach to describe CaCO3 decarbonation in non-steady thermal conditions Estupinan Donoso, Alvaro Antonio ; Peters, Bernhard ; et al in Powder Technology (2015), 275 In cement production, direct measurements of thermal and chemical variables are often unfeasible as a consequence of aggressive environments, moving parts and physical inaccessibility, and therefore ... [more ▼] In cement production, direct measurements of thermal and chemical variables are often unfeasible as a consequence of aggressive environments, moving parts and physical inaccessibility, and therefore prediction models are essential tools in these types of industrial applications. This article addresses the problem of the numerical prediction of the CaCO3 calcination process, which is the first and the most energy expensive process in clinker production. This study was conducted using the Extended Discrete Element Method (XDEM), a framework which allows a Eulerian approach for the gas phase to be combined with a Lagrange one for the powder phase. A detailed validation of the numerical model was performed by comparison to non-isothermal TG curves for mass loss during the CaCO3 decarbonation process. The complex three-dimensional predictions for solid and gas phases are believed to represent a first step towards a new insight into the cement production process. Thus, the high accuracy and detailed description of the problem addressed, serve as a basis to assess the uncertainty of more simplified models such as those used in soft sensors. [less ▲] Detailed reference viewed: 95 (14 UL)XDEM Used for Predicting Tungsten-Oxide Reduction Estupinan Donoso, Alvaro Antonio ; Peters, Bernhard Scientific Conference (2015, April 27) Detailed reference viewed: 88 (8 UL)Performance Evaluation of the XDEM framework on the OpenStack Cloud Computing Middleware Besseron, Xavier ; Plugaru, Valentin ; Mahmoudi, Amir Houshang et al in Proceedings of the Fourth International Conference on Parallel, Distributed, Grid and Cloud Computing for Engineering (2015, February) As Cloud Computing services become ever more prominent, it appears necessary to assess the efficiency of these solutions. This paper presents a performance evaluation of the OpenStack Cloud Computing ... [more ▼] As Cloud Computing services become ever more prominent, it appears necessary to assess the efficiency of these solutions. This paper presents a performance evaluation of the OpenStack Cloud Computing middleware using our XDEM application simulating the pyrolysis of biomass as a benchmark. We propose a systematic study based on a fully automated benchmarking framework to evaluate 3 different configurations: Native (i.e. no virtualization), OpenStack with KVM and XEN hypervisors. Our approach features the following advantages: real user application, the fair comparison using the same hardware, the large scale distributed execution, while fully automated and reproducible. Experiments has been run on two different clusters, using up to 432 cores. Results show a moderate overhead for sequential execution and a significant penalty for distributed execution under the Cloud middleware. The overhead on multiple nodes is between 10% and 30% for OpenStack/KVM and 30% and 60% for OpenStack/XEN. [less ▲] Detailed reference viewed: 231 (48 UL)ROS homeostasis in a dynamic model: How to save PD neuron? Kolodkin, Alexey ; Ignatenko, Andrew ; et al Poster (2014, December) Detailed reference viewed: 90 (14 UL) |
||