Event program

Thursday, 5/27/2010 9:00 AM - 5:30 PM, Galija, Hotel Admiral, Opatija 9:00 AM - 9:15 AM M. Riedel (Forschungszentrum Juelich GmbH, Juelich, Germany), A. Streit, T. Lippert (Juelich Supercomputing Centre, Juelich, Germany), F. Wolf (RWTH Aachen University, Aachen, Germany), D. Kranzlmueller (Ludwig Maximilians University Munich, Munich, Germany) Towards Individually Formed Computing Infrastructures with High Throughput and High Performance Computing Resources of Large-scale Grid and e-Science Infrastructures  The interoperability of e-Science infrastructures like DEISA and EGEE is an increasing demand for a wide variety of cross-Grid applications (EUFORIA, VPH, WISDOM, etc.), but interoperability based on common open standards adopted by Grid middleware systems is only starting to emerge on computing infrastructures and are not broadly provided today. In earlier work we have shown how refined open standards form a reference model, which is based on careful academic analysis of lessons learned obtained from production cross-Grid applications that require access to both, High Throughput Computing (HTC) resources as well as High Performance Computing (HPC) resources. This paper provides more insights in several concepts of this model with a particular focus on realizing interoperability with the so-called 'plumbings approach' in order to satisfy the different security demands of e-scientists and their cross-Grid applications. Based on lessons learned over years gained with initial production interoperability setups and experimental interoperability work between production Grids such as EGEE, DEISA, and NorduGrid, we illustrate how common open Grid standards (e.g. OGSA-BES, JSDL, and GLUE2) jointly together with the plumbings concept contribute to the vision of a network of interoperable Grid services. This network of services in turn overcomes the limitation of purely based HPC-driven infrastructures (i.e. DEISA, TeraGrid, PRACE, etc.) or HTC-driven Grids (EGEE, OSG, etc.) by enabling individually formed computing infrastructures that satisfy the needs of e-scientists with HPC and HTC resources today. 9:15 AM - 9:30 AM Ž. Jeričević (TEHNIČKI FAKULTET, RIJEKA, Croatia), I. Kožar, T. Pecak (Faculty of Civil Engineering, Rijeka, Croatia) Modeling of Linear Algebra System with the Fourier Transform  Solving dense linear system requires N3 order of magnitude operations, where N is a dimension of a square matrix. We developed procedures for modeling and computation of approximate solutions for problems which are too large to be solved routinely and for which an approximate solution is acceptable. Using the ideas from signal processing, the original dense system is Fourier transformed and then pruned off by removing the insignificant frequencies in accord with the rules of matrix algebra. Use of the Fourier transform can be understood as a preconditioning after which significant and spurious components of the system can be separated. This approach reduces the original system size. and leads to a fast, approximate solutions of eigenvalue or linear system of equations problems. Formal transformation of the original system is done through the use of Fourier transform matrices for direct and inverse Fourier transform. In order to increase flexibility of this approach, the partitioning of input matrix was also tested. The partitioning allows the use of mixed approach where part of the matrix is transformed and reduced in size, while other part is kept intact and in original domain. The examples analyzed were from theoretical chemistry and structural engineering calculations. The quality of solutions for practical problems is demonstrated. 9:30 AM - 9:45 AM M. Cankar (Xlab d.o.o, Ljubljana, Slovenia), R. Trobec (Jozef Stefan Institute, Ljubljana, Slovenia) Experimental Evaluation of Grid5000 Performance in the Solution of PDE  The solution of Partial Differential Equations are an extensive and calculation-demanding applications that require high performance numerical computing and fast communication channels. Grids are often not the ideal candidates for the solution of such problems. We evaluated experimentally the performances of Grid5000 in the solution of a diffusion equation using finite difference methods and MPI communication library. Measurements of the point-to-point communication indicate a significant variation in communication speed between grid computers. There are also differences in the numerical performances of the grid computers, either because of different hardware and software resources, or because of sharing the computing power by other users. Both variations could be time dependent. It follows that a quite complex load balancing policy is needed for compensating the differences in the numerical and communication performances and identifying the optimal set of cooperating computers. 9:45 AM - 10:00 AM Ž. Jeričević (TEHNIČKI FAKULTET, RIJEKA, Croatia) In-Depth Analysis of Seismic Algorithm for Surface Multiple Attenuation  In-Depth Analysis of Seismic Algorithm for Surface Multiple Attenuation 10:00 AM - 10:30 AM Pauza za kavu  10:30 AM - 10:45 AM M. van Rijswick (Ulm University, Ulm, Germany), S. Gerhold, T. Schmitt, P. Schulthess (Ulm University, Institute of Distributed Systems, Ulm, Germany) Emulation of a Software-Managed TLB on x86-64 Platforms  In contrast to MIPS and Alpha, the x86-64 platform enforces a specific layout of the page table memory structures. We present a way to emulate the behavior of a MIPS TLB-miss exception on commodity x86-64 hardware. This offers new possibilities to operating system designers as they can freely choose the best memory data structures suitable for the specific operating system purposes. We prove that a page table structure consisting of 6 tables with 4 KB each is sufficient and minimal for this emulation. Additionally we describe a prototype implementation and present measurement results. 10:45 AM - 11:00 AM S. Ivanovska (Institute for Parallel Processing-Bulgarian Academy of Sciences, Sofia, Bulgaria), A. Karaivanova, S. Ivanovska (Institute for Parallel Processing - Bulgarian Academy of Sciences, Sofia, Bulgaria) Matrix computations using quasi-Monte Carlo with scrambling  Quasi-Monte Carlo methods are powerful tools for accelerating the convergence of ubiquitous MCMs. For problems in linear algebra, it is possible to accelerate the convergence of well-known Monte Carlo methods using quasirandom sequences. Moreover, quasi-Monte Carlo methods give smoother convergence with increasing length of the walks which is very important for computing the eigenvalues. In the same time MCMs and QMCMs have the same computational complexity. The disadvantage of quasi-Monte Carlo is the lack of practical error estimates due to the fact that the rigorous error bounds, provided via the Koksma-Hlawka are very hard to utilize. This disadvantage can be overcome by scrambling of the used sequence. Scrambling also gives a natural way to parallelize the streams. In this paper we study matrix-vector computations using scrambled sequences on the grid. 11:00 AM - 11:15 AM D. Davidović (, , ), K. Skala (Rudjer Boskovic Institute, Zagreb, Croatia) Implementation of the WRF-ARW Prognostic Model on the Grid  The numerical weather prediction models demand large execution time and resource allocation; therefore we need large parallel computation facilities, like the grid systems, in order to execute them. In this work the implementation of the Advanced Weather Research and Forecasting (WRF-ARW) model on the SEEGRID-SCI grid infrastructure is presented. The goal of porting the model to the grid is to get more accurate and detailed forecast with operational speed-up on high-resolution model grids. The results and the application of the WRF-ARW model on the Grid will be shown in the example of the bora winds simulations over the Adriatic coast. We have achieved significant progress in accelerating the model execution and simplicity of starting jobs on the grid. The weather forecast is now more accurate due to the increased grid resolution, while the execution time is significantly decreased. In our researches we have found that the benefits are not only in the reduced execution time (up to 30% for large enough number of CPUs) but also in the simultaneous runs of the same model. This is especially important in research activities when many similar jobs can be executed simultaneously on different grid computing elements that results in enormous speedup. We have achieved a great success in porting WRF-ARW model to the grid and improvements gained through deployment of the model increases the support and feasibility of new projects in the area of the SEE-GRID-SCI infrastructure and Earth Science community. Implementation of the WRF-ARW model on the grid infrastructure, increasing scientific research in regional scale particularly in the field of ecology. 11:15 AM - 11:30 AM N. Kaemmer, P. Schmidt, T. Schmitt, P. Schulthess (Ulm University, Ulm, Germany) Garbage Collection Strategies For Multiconsistent Cluster Operating Systems  Garbage Collection is a crucial component of memory management facilities in operating systems, application environments and within applications. Modern systems provide automatic garbage collection thus alleviating the programmer from tracking garbage manually and avoiding incorrect usage. In this paper we present different garbage collection strategies for Rainbow OS, a typesafe distributed operating system for PC-clusters with multiconsistent transactional memory. In addition to known garbage collection schemes such as reference counting or reference tracking we present reverse reference tracking stratehgies using type system information like object dependencies and class relations to detect unused objects in memory. Additionally we describe a possibility to detect garbage using system checkpoints stored by a checkpointing facility outside the transactional distributed memory. 11:30 AM - 12:00 PM Pauza za kavu  12:00 PM - 12:15 PM A. Pajovic (Republicki hidrometeorološki zavod, Banja Luka, Bosnia and Herzegovina), I. Hadzismailovic (Federal hydrometeorological institute, Sarajevo, Bosnia and Herzegovina), D. Davidovic (Rudjer Boskovic Institute, Zagreb, Croatia) Using the grid infrastructure in weather prediction: the WRF-ARW model   In this work the application of the Advanced Weather Research and Forecasting (WRF-ARW) model on the Grid infrastructure will be presented. Here we will present how, once deployed to the grid infrastructure, the WRF-ARW model can be used and gain benefit for weather prediction and research purposes. The application of the model will be shown through the examples of the operational weather forecast and the foehn effects simulations that occurred over Banja Luka in Bosna and Hercegovina (BIH). Using WRF-ARW model and Grid infrastructure is very important to achieve a good weather forecast for small and large regions, but also it’s better to understand how weather behaves in very complex situation or weather conditions. These conditions can be more complicated and depend on local terrain. The weather forecast is often wrong or mistaken for local region due to mountain, the other local terrain conditions or global synoptic situation. One of the most important meteorological events which show us these mistakes is increasing of air temperature caused by well known phenomenon “Foehn effect”. Foehn is a type of dry down-slope wind which occurs in the lee of a mountain range. As a consequence of the different adiabatic lapse rates of moist and dry air, the air on the leeward slopes becomes warmer than equivalent elevations on the windward slope. The aim of this work is to present the increasing of temperature using WRF-ARW model in high resolution. Also, it will be shown how local extreme situation, like as thunderstorm or hailstorm in Banja Luka has occurred last summer, without a clear weather forecast outputs. Foehn effect, WRF-ARW model, Grid infrastructure, weather conditions Aleksandar Pajović, Republic Hydrometeorological Institute, Banja Luka, aleksandar.pajovic@meteo-rs.com Ibrahim Hadžismailović, Federal Hydrometeorological Institute, Sarajevo, ihadzismailovic@etf.unsa.ba Davor Davidović, Ruđer Bošković Institute, Zagreb, davor.davidovic@irb.hr 12:15 PM - 12:30 PM A. Rashkovska, R. Trobec, M. Depolli, V. Avbelj (Jožef Stefan Institute, Ljubljana, Slovenia) Time and frequency analysis of non-uniform sampling  The spectrum of non-uniformly sampled signals is often needed for analysis of signals obtained from network of distributed sensors. In this paper, we analyze a method for retaining the spectrum of this type of signals through interpolation, then re-sampling with uniform rate and performing Fourier transform at the end. We discuss this method based on the deviation of the non-uniform sample rate and the interpolation type. Then, we examine the methodology on a non-uniformly sampled sine wave and a real ECG signal obtained from a Wireless Sensor Network (WSN). 12:30 PM - 12:45 PM J. Matkovic (JP Elektroprivreda HZ-HB d.d. Mostar, Mostar, Bosnia and Herzegovina), K. Fertalj (FER, Sveučilište u Zagrebu, Zagreb, Croatia) Comparative Analysis of Web Services and Web Service Development Technologies  This paper focuses on ontology and taxonomy of services. Service is a paradigm frequently mentioned with the genesis of idea about Service Oriented Architectures - SOA. SOA is still discussed even though the idea about such implementation of services is not new. In the last couple of years, SOA has become attractive, perspective and promising even though there are lots of obstacles in implementation of SOA concepts in practice. Inadequacy of SOA lies in a fact that there are lots of obstacles connected to defining of clear SOA standards. Whereas main structural unit of SOA architecture is service, it is quite clear, that precise definition of service is something of big importance in order to go further in synthesis of SOA elements: Enterprise Service Bus, communication protocols, Repository for Service Discovery and many others. The first purpose of this paper is to define the function of services as autonomous software drives, developed to support business processes, and executed when needed. Afterwards paper discusses several platforms for Web service development, offered by the most relevant vendors of software solutions. Discussion explains the ways in which vendors conceived Web service paradigm and implemented it in their solutions, with insight into their advantages and disadvantages. 12:45 PM - 1:00 PM P. Taler (Sveuciliste u Osijeku, Odjel za matematiku, Osijek, Croatia), D. Matijevic, P. Taler (Odjel za matematiku, Sveuciliste u Osijeku, Osijek, Croatia), G. Martinović (Elektrotehnički fakultet, Sveučilište u Osijeku, Osijek, Croatia) DISTRIBUTER – The Distributed System for Efficient Execution of Parallel Programs  We developed the distributed system “Distributer” in order to make use of millions of unused CPU cycles inside the LAN. The system is based on server-client architecture and the interaction with the system is implemented via web interface on the server side. Clients connect to server and periodically update their status based on which server distribute and send jobs and at the end collect results. Our system is easy-to-use and easy-to-install and has the simple centralized management of applications needed on client side for jobs execution. We keep all applications on one central repository that is shared to all clients inside LAN with the help of network file system protocol (NFS). The current implementation of Distributer accommodates several different applications (C, C++, FORTRAN, Python or some commercial products, such as Wolfram’s Mathematica and MathWorks’ Matlab) and it provides flexible and scalable mechanism for distributing data. Jobs are submitted in a form of a single input file (e.g. input file for some of supported applications or a coordinating script). 3:00 PM - 3:15 PM Q. Pan, G. Reitmayr, E. Rosten, T. Drummond (University of Cambridge, Cambridge, United Kingdom) Rapid 3D modelling from live video  ProFORMA is a system capable of real-time 3D reconstruction of textured objects rotated by a user's hand. Partial models are rapidly generated from the live video and displayed to the user, as well as used by the system to robustly track the object's motion. This talk will introduce some of the techniques used to enable the creation of this real-time modelling system as well as look at what we can expect from such systems in the future. 3:15 PM - 3:30 PM C. Mocan (Technical University of Cluj-Napoca , Cluj-Napoca, Romania), D. Gorgan (Technical University of Cluj-Napoca, Cluj-Napoca, Romania) Cluster Based Modeling and Graphical Visualization of Interactive Large Spatial Data  Rendering and interacting with high-resolution geographical data sets and complex models of virtual geographical space involves high power computation resources in networking environments. The solution of integrating graphics rendering engine applications in cluster based architecture and Grid infrastructure is the main concern of the research reported by this paper. The performance of load balancing is evaluated for various cluster configurations by considering different combinations of distributed rendering algorithms over the graphics cluster and spatial data models. The research studies and experiments as well the flexibility related with the mesh data formats accepted by the graphics renderer, user interaction techniques with complex scenes in the context of graphics cluster rendering, and solutions for data streaming and unit frame encoding. A few use cases of visualization of the virtual geographical model exemplify the achievements. 3:30 PM - 3:45 PM V. Bojović (Institut Ruđer Bošković, Zagreb, Croatia), I. Grubišić (IRB, Zagreb, Croatia), B. Lučić, K. Skala (Institut Ruđer Bošković, Zagreb, Croatia) Analyser and Viewer of Protein Inter-Residue Contacts  This service is made in order to achieve more efficient insight into the nature of hydrophobic or hydrophilic forces in proteins. It enables to visualize details and specificity in contacts between amino acid residues in proteins. In addition, it is possible to estimate and calculate contributions of inter-residue contacts depending on selected physical and chemical properties of amino acid residues. Choosing one amino acid residue, distances between its sidechain and sidechains of all amino acid residues within the sphere of choosen radius *r * can be visualized. Additionally, by defining the functional form of physical contribution of inter-residue contacts to protein folding, contributions of all contacts can be summarized over all resodues for selected protein. Current version of database included in the server contains only 42 proteins crucial for performing research on modeling of protein folding rates, but it will be enlarged by new sequences. Visualization is available in png, povray, and vrml format, and the PDB format output is also available. 3:45 PM - 4:00 PM K. Antlej (IB-PROCADD d.o.o., Ljubljana, Slovenia), B. Reberšek (Marmor Hotavlje d.d., Hotavlje, Slovenia), D. Cugelj (Gorenje Design Studio d.o.o., Ljubljana, Slovenia), J. Jelenc (Iskra MEDICAL d.o.o., Ljubljana, Slovenia) Use of 3D technologies in new product development  Problems of conventional new product development (NPD) consist of long standing procedures, unrepeatability of processes that are also difficult to track, inaccurate realizations, poor communication between subjects, and large share of human errors. 3D technologies enable implementation of DSSP (Digital Shape Sampling and Processing), RP/RM/RT (Rapid Prototyping/Rapid Manufacturing/Rapid Tooling) and other related digital applications into entire PLM (Product Lifecycle Management) from product conception, through development, to marketing and are therefore successfully solving above mentioned problems. Introducing state-of-the-art 3D technologies into NPD and PLM is, because of its parametric and digital form, therefore necessary. This paper analyses the use of 3D technologies in industry and describes the most recent projects (processes) developed by IB-PROCADD company in cooperation with its partners. 4:00 PM - 4:30 PM Pauza za kavu  4:30 PM - 4:45 PM M. Stanić (Grafički fakultet, Zagreb, Croatia), B. Lozo (Faculty of Graphic Arts, University of Zagreb, Zagreb, Croatia) Colour and Permanence Issues in 3D Ink-jet Printing  3D ink-jet printing is a powder-based rapid prototyping (RP) technique. It is based on the conventional ink-jet printing method. The digital 3D model can be obtained by e.g. 3D scanning or by designing in 3D modeling software. The model is virtually sliced in layers, which are being printed by consecutive layers (cross-cuts) of basic materials, powder and binders. As the 3D printing by ZCorporation, described here, is capable of printing in “full” colour, it uses both clear and coloured binders. Their current systems use either a combination of CMY or CMYK coloured binders and the colour is achieved similarly as in conventional colour ink-jet printing. In this work, the basis of colour reproduction and colour measurements methodology, as well as the colour permanence or colour stability, are described and discussed. The issues which are explained include design of 3D colour test charts (generic and custom developed) as colour visualization aids, colour appearance and colour variations in 3D prints, colour measurements of 3D prints and colour permanence of 3D prints. The work further explains the current possibilities of colour 3D ink-jet printing and defines the importance of colour monitoring and control, as well as the colour permanence issues, in 3D printing. 4:45 PM - 5:00 PM K. Žiha (FAKULTET STROJARSTVA I BRODOGRADNJE, ZAGREB, Croatia), K. Skala (IRB, Zagreb, Croatia), D. Ogresta (Noble&Denton, Dubrovnik, Croatia), B. Marunica (FSB, Zagreb, Croatia) Virtual Modeling in Shipbuilding  The introduction of this paper copes with presentations of ship constructions and 3D visualizations. Next, the paper considers the needs, methods and applications of virtual modeling in shipbuilding with emphasis on the application of Virtual Modeling Language – supporting the VRML format. Finally, the paper presents the experiences in conversion of small and bigger, partial and integral three dimensional models of ships into virtual models for active usage in real time and in network environment. The work demonstrates that the creation of big virtual models in shipbuilding is feasible but requires a number of accommodations and optimization of the model size that affects the clarity of the model details as yet. The GRID application may contribute to photorealistic visualization in network environment. The conclusion supports the thesis that in the next future virtual models of complex ship structures will be the prevailing methods of presentation, often as parts of more far-reaching virtual laboratories. 5:00 PM - 5:15 PM M. Fifik (Technical University of Košice, Kosice, Slovakia), J. Turán, Ľ. Ovseník (Technical University of Košice, Košice, Slovakia) Real Time Recognition System for Traffic Sign Detection and Classification  This paper contains topic from Machine Vision field – detection of traffic signs, captured in moving cars. Detailed recognition system for real time detection is described. This system include following steps: input image is converted in to the HSV color space or RGB color space. After this three binary maps are created. Every binary map is cleared from irrelevant information. Together with image clearing, boundaries are from possible region of interest detected (ROI). When is ROI detected, shape detection is applied on every ROI. Shape detection is done by pattern matching. Then information about passed ROI(s) is used in block invariant feature extraction, where Trace Transform is used. All information from preprocessing and invariant feature extraction block is used in sign type classification. Potential of this system is discussed. 5:15 PM - 5:30 PM I. Kožar (GRAĐEVINSKI FAKULTET SVEUČILIŠTA, RIJEKA, Croatia) Visualization of complex civil engineering structures  Civil engineering is a broad discipline so that visualization in it includes visualization of natural phenomena like terrain and watercourses, as well as visualization of large man-made structures and transportation facilities. There is also visualization of mathematical models and results. This work presents visualization of terrain and corresponding watercourse from geodesic data. Further are presented finite element computer models of large engineering structures. There is a short description of microplane material model used in the analysis of those large quasi-brittle structures as well as some aspects of presentation of the results. In the following is visualization of some contact problems relevant in civil and structural engineering. At the end is a short presentation of author’s original method for visualization of implicit mathematical functions.