The Leading Source for Global News and Information Covering the Ecosystem of High Productivity Computing
November 02, 2007
ORNL Tests Long-Haul InfiniBand over UltraScience Net
RENO, Nev., Oct. 31 -- Obsidian, the leader in InfiniBand range-extension, is providing Oak Ridge National Labs (ORNL) with a means of testing long-haul InfiniBand over UltraScience NET (USN). Dr. Nageswara Rao, USN's co-Principal Investigator, said, "We are interested in technologies that will help us move increasingly large datasets between Department of Energy (DOE) labs and facilities. The Longbows on USN are posting remarkable performance numbers and offer a very attractive alternative to TCP/IP."
The DOE operates many of the world's largest "capability" supercomputers, producing and consuming data files that are growing exponentially in size as ever larger machines are deployed -- increased data capacities enable larger, higher-fidelity physics simulations. However, this growth trend presents a critical problem to the DOE: "When supercomputer users need to move super-sized files, shared-infrastructure TCP/IP-based networks are unable to handle the bandwidth overload," remarked Dr. Rao. "Clustering techniques for processors and storage have allowed the computing machinery to scale, but the conventional approaches to networking them together have not kept pace."
ORNL has been evaluating Obsidian Longbow XR devices since September 2006, running a variety of tests over USN, which is well suited as a next-generation networks testbed. Comprising a pair of OC-192 SONET channels between Oak Ridge, Atlanta, Chicago, Seattle and Sunnyvale, it can be configured to provide loops of varying lengths. Results have been highly encouraging, according to Dr. Rao: "Longbow XRs configure trivially easily, simply appearing as two-port InfiniBand switches that completely hide the optical Wide Area Network (WAN) from the InfiniBand servers and storage. Using InfiniBand's Remote Direct Memory Access (RDMA), we found we could stream data across all 8,600 miles of USN with an essentially flat bandwidth-distance curve, right out of the box. To see this kind of plug-and-play performance with no in-situ tuning, and independent of distance, is in striking contrast to normal network performance."
Dr. David Southwell, President and CEO of Obsidian Research, said, "We are most pleased with Dr. Rao's results using Longbow XRs across USN, and look forward to working with the DOE in the future." Work is already underway to integrate the InfiniBand-enabled Lustre file system to provide file level native InfiniBand transport between DOE locations at very high speeds.
Dr. Rao is providing live demonstrations of this capability across USN between ORNL in Tennessee and Reno during SC07 in the Oak Ridge National Labs booth.
About Obsidian
Obsidian Research Corporation and the Obsidian Longbow LP are the developers of Longbow, a series of InfiniBand range extension products. Longbow technology allows an InfiniBand fabric, normally a short-range network used in high-performance computing, to be extended via optical fiber over varying distances. Longbow connects across Campus, Metro or Global networks to offer unparalleled high-bandwidth, low-latency access to InfiniBand compute and storage resources. For more information, see http://www.obsidianresearch.com/.
About UltraScience Net
The next generation DOE large-scale science applications have requirements that will drive extreme networking, in terms of sheer high throughputs as well as dynamically stable connections. The UltraScience Net is (USN) an experimental research testbed funded by DOE Office of Science to enable the development of hybrid optical networking and associated technologies to meet the unprecedented demands of large-scale science applications. For more information, see http://www.csm.ornl.gov/ultranet/.
-----
Source: Obsidian Research Corp.
FREE Download: "Going Parallel - An Implementation Guide"
Breakthrough performance for MATLAB®, Python and other desktop apps... Get 100X speedups, with less than 10% of the development time. Focus is on enabling familiar desktop tools to virtually execute on parallel servers, clusters, and grids.
Podcast: John West talks with John Lee, Appro’s VP of Advanced Technology Solutions at SC08
Learn more about Appro, what they showcased at SC08 this year, and where the company is headed. Also joining the conversation was Graeme Hackland, the IT manager for the Renault F1 team, one of Appro's big HPC customers.
RSS Feeds
Feeds by Topic
- Applications
- Developer Tools
- Interconnects
- Middleware
- Networks
- Processors
- Storage
- Systems
- Visualization
Feeds by Industry
Feeds by Content Type
Feature Articles
A Framework for Interfacing Configurable Hardware Accelerators
For all the benefits claimed by hardware acceleration, from exponential performance improvements to massive power and space savings, most of these benefits focus on what can be accomplished with little detail on how to accomplish it. Hardware acceleration always seems to have the implied acronym DIY (Do It Yourself).
Read More...
Santa's HPC Woes
You thought you had it hard. What if all of your business, with a worldwide customer base, occurred in the midnight hours of a single day towards the end of December?
Read More...
Baby, It's Cold Outside...Let's Calculate Something
Argonne National Labs is part of the Department of Energy, so it's not exactly surprising to learn that they are actively looking for ways to reduce energy use. But using Chicago's cold winters to save $25,000 a month on cooling costs for the supers in their Leadership Computing Facility is, well, cool.
Read More...
Top Headlines
How 100 Gigabit Ethernet Markets Will Differ from 10
Jan 07 | Seeking Alpha | The advent of 10 Gigabit Ethernet created major opportunities for equipment startups, but this may not be the case for the future 40 and 100 Gigabit Ethernet markets. Read more...
Virtual Product Development Gaining Traction
Jan 05 | Design News | While manufacturers may never be able to completely avoid physical prototyping, best-in-class players are continuing to expand their use of virtual product development. Read more...
C-DAC Initiatives for NE Region
Jan 05 | CXOtoday.com | The Centre for Development of Advanced Computing (C-DAC) has inaugurated its PARAM Sheersh Supercomputing Facility at North Eastern Hill University in India. Read more...
China-Made Supercomputer to be Installed in Shanghai
Jan 02 | People's Daily Online | A 200 teraflop Chinese supercomputer is scheduled to be installed in Shanghai in April. Read more...
Finding Huge Potential in Rocks’ Tiniest Cracks
Dec 31 | Houston Chronicle | A Houston-based company is using supercomputing power to reveal how gas and oil flows through underground rocks. Read more...
Featured Whitepapers
Star-P: Parallel Computing Without Parallel Programming
Dec 15 | | Engineers, scientists, and other domain experts depend on the productivity enabled by very high-level language (VHLL) tools like MATLAB® and Python. However, as datasets grow larger and programs get more sophisticated, ordinary desktop computers can no longer keep up. The paper explores how to run VHLL programs on high-performance platforms without low-level reprogramming. Work with large datasets and complex algorithms without sacrificing ease of use or reducing productivity.
Virtual Wind Tunnel: Building the ING Renault F1 Team Aerodynamics Research Supercomputer
Dec 02 | | Learn about Appro’s experience providing a supercomputer for the use of Computational Fluid Dynamics at the ING Renault F1 Team’s technical centre in the United Kingdom. The Appro Xtreme-X™ Supercomputer provides a balanced HPC architecture for scalable performance, power and cooling efficiency, and high availability computing in a redundant and manageable framework.
Multimedia
Webcast: HPC Development Solutions: Sun Studio & Sun HPC ClusterTools
Sun Studio Compilers and Tools and Sun HPC ClusterTools allow you to create high performance parallel applications for OpenSolaris, Solaris and Linux. Sun Studio Express 11/08 includes MPI performance analysis capabilities and full OpenMP 3.0 compiler support. Learn about all this and the latest in Sun HPC ClusterTools 8.1.
Webcast: Dell Expands HPC Access and Adoption with Intel Cluster Ready Program
Source: Addison Snell, GM/VP, Tabor Research; sponsored by Dell
Many organizations that could benefit from the use of HPC clusters find that it is complicated to get the systems up and running because of limited IT resources or the complexities of the clusters themselves. Learn how the Intel Cluster Ready program, for which Dell was an original partner, seeks to address this challenge for entry level and mid-range HPC users.
Video White Paper: Architecting a Better Network Storage Solution
BlueArc's Titan architecture represents an evolutionary step in file servers by creating a hardware-based file system that can scale bandwidth, IOPS, and overall data capacity well beyond conventional software-based devices. With its ability to virtualize a massive storage pool of up to four usable petabytes of tiered storage, Titan can scale with growing data requirements, offering a competitive advantage for businesses, researchers, or other enterprises seeking to better manage data growth while still ensuring optimal performance.
Newsletters
Stay informed! Subscribe to HPCwire email Newsletters.
HPC Job Bank
Featured Events
-
January 13-14, 2009
1st European Conference on Software Services and SOKU technologies
Brussels Belguim
-
February 24-27, 2009
TradeTech Architecture 2009
London England
-
February 25-27, 2009
Adopting Integrated R&D Informatics Systems
San Francisco , CA
United States
-
March 11-13, 2009
HPC Horizons Summit
Lake Buena Vista , Florida
USA
-
April 13-16, 2009
Symposium 2009: LEAN, CLEAN, and GREEN
New York , NY
USA
-
April 27-29, 2009
Bio-IT World Conference & Expo
Boston , MA
United States