SC06 Topics

High Performance Computing for Nano-science and Technology

Professor Jun Ni will be discussing High Performance Computing for Nano-science and Technology. HPCNano06 is the second workshop that will have conducted in conjunction with IEEE/ACM Supercomputing 2006 (SC|06). The workshop will be guided by the SC2006’s Workshop Committee and planned and executed by the workshop program committee. We hope to attract people from diverse science and engineering disciplines, nationally and internationally, to attend the workshop, present their research results, share their experiences and ideas, and plan future collaborations.Further information can be found in the workshop’s Web site at:
http://www.uiowa.edu/~nano/HPCNano06/

Nanotechnology is an exciting field with many potential applications. Its impact is already being felt in materials, engineering, electronics, medicine, and other disciplines. Current research in nanotechnology requires multi-disciplinary knowledge, not only in sciences and engineering but also in high performance computing (HPC) technology. Many nano-science explorations rely on mature, efficient HPC and computational algorithms, practical and reliable numerical methods, and large-scale computing systems. This workshop offers academic researchers, developers, and practitioners an opportunity to discuss various aspects of HPC-related computational methods and problem solving techniques for nano-science and technology research.

GROW - Grid Research and Education Group at Iowa

Mission

The Grid Research and educatiOn group @ IoWa (GROW) was established at the beginning of year 2003 to conduct Grid computing research, education, and outreach by integrating The University of Iowa and the state of Iowa computational resources. The vision for GROW is to advance computational science and engineering by broadening and strengthening connections between Grid and high performance computing community and domain sciences and engineering communities. The main strategy guided by this vision is to create a shared Grid-based campus and state cyberinfrastructure that interoperates with evolving national (e.g.,the Open Science Grid and TeraGrid) and global Grid-based cyberinfrastructure (e.g., Large Hadron Collider Computing Grid).

People

GROW members include a number of research and education staff, faculty, graduates, and undergraduates with diverse disciplinary expertise such as Computer Science, Engineering, Geography, Physics, Radiology, and Statistics. Dr. Shaowen Wang, a research scientist of UI ITS AT-RS and an adjunct assistant professor in UI Department of Geography, is the founder and manager of GROW.

Projects

Current research and development projects range from an intra-campus Grid (i.e., the HawkGrid) that interfaces with evolving national and global Grid-based cyberinfrastructure, a proof-of-concept physics Grid Tier-2 center, Grid portal-based user-level services (e.g., resource monitoring, discovery, and scheduling) in the context of problem solving environments, application-specific Grid middleware for computationally intensive problem solving, and distributed computing architectures for desktop Grids, the GROW virtual organization in the Open Science Grid, to the Open Science Grid Generic Information Provider.

MIP: Modular Information Provider

The Modular Information Provider (MIP) is developed by GROW to manage and automatically aggregate information sources for Grid information services. MIP tackles the challenge of mapping the information from a number of sources to be handled by a few information services with minimal human intervention. MIP addresses such mapping through a modular approach through which it can be deployed to achieve the interoperability among multiple Grids. MIP can be customized in a straightforward way to a specific Grid environment, and it also supports flexible information schema. Current MIP implementation is based on Globus MDS4 and the XML version of GLUE Schema 1.2.

The design of the Modular Information Provider (MIP) aims to address the shortcomings that exist in Grid information providers (e.g., Generic Information Provider) as well as to support the web service-based Grid information services. Our modular approach has been developed to ease maintenance and management of Grid information systems by automatically pulling information sources together and filling appropriate pieces of information into Grid information services based on information schemas. This approach minimizes memory requirements as it allows dynamically loading only necessary modules that are customizable to the requirements from a particular Grid resource.

GISolve: A Grid-based Problem Solving Environment for Computationally Intensive Geo. Info. Analysis

Geographic information is collected to support scientific investigations and decision-making in a wide variety of application domains (e.g., environment science, transportation, public health, and business). Enormous computational resources are needed to store and manage geographic information that is collected for such diverse purposes, and to conduct computationally intensive geographic information analysis. This type of analysis is a main focus of Geographic Information Science (GIScience), an interdisciplinary field involving geography and other social sciences, computer science, geodesy, and information sciences to study generic issues about the development and use of geographic information systems (GIS) technologies. The purpose of this project is to develop a TeraGrid Science Gateway toolkit for GIScience. Our gateway toolkit provides user-friendly capabilities for performing geographic information analysis using computational Grids, and help non-technical users (such as from social sciences) directly benefit from accessing Grid capabilities.

SOG - Self Organized Grouping Method for Grid Resource Discovery

The Bootable Cluster CD and LittleFe

The Bootable Cluster CD (http://bccd.cs.uni.edu) and LittleFe (http://LittleFe.net) are a practical low-cost two-tiered cyberinfrastructure approach we have developed. The BCCD is a live CD that transforms an x86 or PowerPC based lab into an ad-hoc computational cluster for educational purposes. LittleFe is an inexpensive, portable, 6+ node computational cluster (running a liberated BCCD) whose primary applications are education, outreach, and training. LittleFe is visceral; students do and see the science, e.g. a protein folding, and being visualized, in real-time. This makes a very engaging platform for the xbox generation.

The principle goal of BCCD/LittleFe is to reduce the friction associated with teaching parallel and distributed computing, Grid computing, and computational science to a range of audiences: undergraduate science faculty, undergraduate students, 9-12 science teachers, and 9-12 students. For example, with the combination of LittleFe and the BCCD's curriculum modules, it only takes 10 minutes to setup and run a live demonstration of molecular dynamics (e.g. protein folding). The simulation is first run locally on LittleFe, on a relatively small scale, and then with a local Grid portal to the TeraGrid for a much larger run.

CT Imaging Applications on a Peer-to-Peer Network