| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
NOTUR2006 – May 11–12, 2006 – Bergen Home - General - Location - Programme ProgrammeLast-minute changes are possible.THURSDAY MAY 11Registration starts at 9.30am. The programme starts at 10.00am with talks on Grand Challenge applications in computational science. This continues in the afternoon and is followed by talks on national and international infrastructure projects. The day finishes ca 5.10pm. A conference dinner is scheduled in the evening at the Bergen Aquarium.
A group picture will be taken just before lunch (12:10). FRIDAY MAY 12The programme starts at 9.00am and includes talks on Grid infrastructures and e-Science, a session with posters from master and PhD students, and a round-table discussion on e-Science. The day finishes ca 4.15pm.
Abstracts Thursday May 11Computational challenges in computational chemistry and Norwegians solutions to improving computational efficiencyKenneth Ruud, University of TromsøIn the talk, I will briefly introduce the equations defining the foundation for molecular quantum chemical calculations. Emphasis will be put on explaining the scaling of the algorithms in use for solving these equations, and how this poses significant challenges for computational chemistry in the future. A few examples from recent research in the Norwegian theoretical chemistry groups aimed at reducing the scaling of these algorithms, or improving the performance of quantum chemical calculations on parallel machines will be given. These examples will be based on research on the quantum chemistry program Dalton, a large Europen, free-of-charge quantum chemistry program, with significant contributions from Norwegian researchers. An outlook on future challenges in quantum chemistry and a few ideas for attacking these problems will also be given. Visualization and computation activities at CMRInge Kristian Eliassen, Christian Michelsen Research, BergenMany of the software projects at Christian Michelsen Research (CMR) have computational related aspects. A wide range of visualization techniques is used both on desktop and in VR environments. Data sources include computational fluid dynamics, seismic data cubes and surveys with multi-frequency echo sounders. Processing methods such as categorization, gridding and various kinds of filtering are performed both real-time and in post-processing mode. This presentation will give an overview of some of the projects at CMR covering broad ranges of application domains, data types and processing methods. Keywords: Visualization, Computation, Data processing, Software development The Evolution of the Supercomputing Market: New Applications, New Technologies, New ApproachesDave Turek, Vice President, Deep Computing IBMThis talk will examine some of the marketplace trends and the technological and economic issues which are leading to innovative approaches to supercomputer design and application. In particular, the presentation will provide some insights into the evolution of the Blue Gene design and consideration of how emerging technologies like the Cell Broadband Engine could be exploited for high performance computing applications. MareNostrum: a system for Grand Challenge applicationsFrancesc Subirada (speaker) and Sergi Girona, Barcelona Supercomputing Center, SpainPresentation of MareNostrum, a system with 4812 processors, interconnected with Myrinet, 9.6 Tbytes memory and 140 Terabytes disk space. MareNostrum is located at the Barcelona Supercomputing Center, where we provide services, but we focus on research on the areas of life science, earth science and IT. MareNostrum is running since June 2004, with an average utilization of 85%. Most relevant is the effort in running Grand Challenge applications, those using an important part of the system, thousands of processors. Grand Challenge users are welcome to the system. The Worldwide Large Hadron Collider Computing Grid from a Nordic perspectiveMichael Grønager, Niels Bohr Institute, Copenhagen, DenmarkThe Large Hadron Collider experiments are on the footstep to begin their huge production of new scientific data, which are expected to have very high impact on out current view of the standard model. The amount of data created is huge - roughly one DVD each second. In order to store and process these vast amount of data a worldwide computation grid is currently under construction. The grid is a collaboration between many countries and three different types of grid middleware: The American Open Science Grid, the European gLite and the Nordic ARC. This talk is an overview of the efforts currently taking place in the Nordic countries to become ready for the data production next year - work which is being coordinated by the Nordic Data Grid Facility. Notur - The Norwegian Infrastructure for Computational ScienceJacko Koster, UNINETT Sigma AS, TrondheimAn overview will be given of the Notur project that provides the national infrastructure for Computational Science in Norway. Since January 2006, Notur is a project within the research program eVITA on eScience and eInfrastructure that was established by the Research Council of Norway. Current status, plans, and future directions for the project will be highlighted. A new 7 TFLOP cluster in NoturBjørn Lindi, NTNU, TrondheimFrom the next allocation period, 2006.2, a new 7 TFLOP cluster will be available in Notur. The cluster is based on IBM Power5+ architecture with IBM Federation technology as interconnect. The new computing resource will almost triple Notur's total computing power. It will be used for operational weather forecast, CFD, construction mechanics, and physics calculations. Besides a presentation of the system, a short how-to-conduct-an-HPC acquisition will be given. The Finnish HPC Infrastructure, the strengths and challenges of the centralized service modelKimmo Koski, Managing Director, CSC, Espoo, FinlandIn Finland, the IT infrastructure for computational science is centralized to a single service center, unlike in other Nordic countries. This has evolved during the last decades through the development of CSC, the Finnish IT center for science. This talk concentrates on the Finnish activities in supporting high performance computing and discusses the strengths and challenges of the centralized model. In order to be successful, emphasis for international collaboration is required. The future infrastructure development plans in Finland and their linkages to international research infrastructure collaboration, such as Nordic and European grid projects, are discussed during the presentation. Abstracts Friday May 12Application Performance on Commodity-class Computers: The Impact of Cluster ArchitectureMartyn F. Guest, Associate Director, Computational Science and Engineering Department, CCLRC Daresbury Laboratory, UKIn many application areas, commodity-based clusters provide an established, cost effective solution to the provision of High Performance Computing. The ability of such clusters to address the differing demands from the fields of Capability (high-end) and Capacity (mid-range) computing is, however, very much a function of both application domain and detailed cluster architecture - node and interconnect. In this presentation we compare the performance of a variety of clusters with current high-end hardware, such as the IBM p5-575 series, the SGI Altix 3700, HP's Superdome and the Cray XT3. We overview performance data from a variety of commodity-based systems (CS1-CS28), featuring IA32, x86-64 and IA64 architectures, coupled to a variety of interconnect technologies - Gbit Ethernet, Myrinet, SCI, Infiniband, Infinipath and Quadrics QSNet. Our focus lies in applications and concentrates on the areas of computational chemistry, computational materials and computational engineering. Based on simple metrics, we consider the performance of variety of codes from each of these disciplines, including NWChem and GAMESS-UK, CPMD, DLPOLY, plus ANGUS and PCHAN. In each case we identify the associated bottlenecks as a function of discipline, and provide a guide to the optimum cluster architecture that might be deployed in a given application domain. DEISA Supercomputing InfrastructureKlaus Lindberg, CSC, Espoo, FinlandDEISA is a consortium of leading national supercomputing centres in Europe. It currently deploys and operates a persistent, production quality, distributed supercomputing environment with continental scope. The purpose of this FP6 funded research infrastructure is to enable scientific discovery across a broad spectrum of science and technology, by enhancing and reinforcing European capabilities in the area of high performance computing. DEISA is structured as a layer on top of the national supercomputing services, and coexists with them. This becomes possible through a deep integration of existing national high-end platforms, tightly coupled by a dedicated network and supported by innovative system and grid software. DEISA provides leading scientific users with transparent access to a European pool of computing resources. The coordinated operation of this environment is tailored to enable new, ground breaking applications in computational sciences. DEISA organizes regular symposiums to the scientific user community. EGEE-II and the Northern Europe Regional Operations CentrePer Öster, PDC, KTH, Stockholm, SwedenEGEE-II and evolution of related Grids in Northern Europe and beyond. Simulating Tsunamis on Parallel ComputersXing Cai, Simula Research Laboratory, OsloOcean-scale simulation of tsunamis presents a computational challenge, not only because millions of degrees of freedom are involved, but also because different physical features must be considered in different regions. For example, dispersion is important at open sea with large depth, whereas nonlinear effects are essential in shallow regions near the shore. To handle the large amount of computation, while treating varying physics throughout an entire ocean, we have adopted a hybrid software framework designed for parallel computers. More specifically, the entire solution domain is decomposed into a set of subdomains, each assigned to a different processor. A subdomain is allowed to be relatively independent, such that it is free to choose its own mathematical model, discretization technique, mesh type/resolution, and even its own software. Coordination and communication between the subdomains are handled by a generic software library implementing the additive Schwarz strategy. This presentation will also report concrete simulation results produced by a hybrid parallel tsunami simulator. eScience - Infrastructure, Theory and Application (eVITA). A Research and Infrastructure Programme, the Research Council of NorwayMorten Dæhlen, Simula Research Laboratory and Chairman eVITA Board"The most scientifically important and economically promising research frontiers in the 21st century will be conquered by those most skilled with advanced computing technologies and computational science applications." (Executive summary, page 1, in Computational Science: Ensuring America's Competitiveness, by the President's Information Technology Advisory Committee, USA, 2005.) eVITA is a research and infrastructure programme designed to address computing- and data-intensive challenges in science, technology and medicine. By promoting research on methodologies, competence development and investment in new eInfrastructure, eVITA will work to ensure that Norwegian research in the eSciences achieves a high international standing, and seek to address important national challenges in the national priority areas of energy and the environment, oceans, food, and health. Mathematics, informatics and statistics, - all disciplines that focus on methodologies - together with a selection of scientific and industrial applications, comprise the main body of the programme's research component. The infrastructure component will ensure adequate and reliable access to computing power and resources for handling large volumes of data for Norwegian researchers and for operational weather forecasting. The eVITA research programme will also provide Norway with an effective tool for mounting a concerted research effort in the eSciences and investment effort in relevant eInfrastructure. Posters Friday May 12The following posters from master and PhD students were accepted for NOTUR2006:
|