Berkeley Lab To Highlight HPC And Networking Expertise at SC99

NOTE: This archived news story is made available as-is. It may contain references to programs, people, and research activities that are no longer active at Berkeley Lab. It may include links to web pages that no longer exist or refer to documents no longer available.

At SC99, the conference for high-performance computing and networking to be held Nov. 13-19 in Portland, Ore., the U.S. Department of Energy's Lawrence Berkeley National Laboratory will feature a display chronicling the 25-year history of the National Energy Research Scientific Computing Center (NERSC) and the Energy Sciences Network (ESnet), as well as new applications for shaping the future of computational science.

As the first national supercomputing center open to the research community, NERSC has served as a model for similar centers across the nation. NERSC today is a leader in providing users with the strongest combination of computing facilities, new tools and technologies, and intellectual resources. ESnet, a backbone network of the Internet, helped pioneer the concept of using phone and satellite links to remotely access supercomputers and today provides high-bandwidth connectivity linking 30 DOE sites around the country. As part of the Berkeley Lab display, various components from supercomputers of the past will be on display -- and attendees who can identify the artifacts accurately will be eligible to win a prize.

Berkeley Lab’s Computing Sciences organization will also showcase a number of applications aimed at making supercomputing applications more powerful, more accessible and more capable. In addition to a dozen interactive technology demonstrations, Berkeley Lab researchers will share their expertise by presenting tutorials and technical papers in sessions open to the nearly 6,000 conference attendees.

Berkeley Lab’s tutorials range from demonstrating how to build cluster computers based on the Linux operating system to looking ahead to “High Performance Computing Facilities for the Next Millennium” to examining the role of high-performance computing in advancing research in molecular biology.

Technical papers presented by Berkeley Lab researchers will cover such topics as the performance of leading-edge supercomputer architectures, analysis of global climate data, new tools for conducting scientific experiments remotely over networks, and applications for analyzing problems such as architectural lighting and studying the movement of underground chemicals.

Technical demonstrations in the Berkeley Lab booth will include:

• A new application that will enable researchers using desktop workstations to generate scientific visualizations from large volumes of data. This new “volume rendering” short-cuts the previous need to have direct access to powerful supercomputers when complex, three-dimensional scientific problems.

 • DeepView, a scalable system for linking experts and facilities from geographically dispersed locations via a wide area network. This “Channel for Distributed Microscopy” offers a listing of available on-line microscopes where users can seamlessly participate in an experiment, acquire expert opinions, collect and process data and store this information in electronic notebooks.

 • On-line access to large scientific datasets using a range of Data Grid technologies. Data Grid nodes at the Berkeley Lab, Argonne National Lab and VA Linux booths on the SC99 exhibit floor and at sites around the country are connected via high-speed transfer mechanisms, allowing SC99 participants to request and track the creation of new dataset replicas and interactively browse dataset contents via Grid-enabled visualization tools.

• Scientific visualization of magnetically confined plasmas in stellarators, which provide important insights in the field of magnetic fusion energy. Such visualizations help improve the design of fusion devices, as well as understanding of plasma physics.

• A prototype “virtual protractor” which allows researchers to measure objects and distances in three-dimensional visualizations. Such models can range from molecules to mile-wide oilfields.