Saturday February 23, 2008 YouTube: TACC Ranger Dedication Ceremony
The Sun HPC Portal brings us this exclusive video of the TACC Ranger dedication ceremony, which took place on February 22, 2008. The event was also covered by the American Statesman in this story.
Honored speakers included:
* William Powers, Jr., President of the University of Texas at Austin
* Daniel E. Atkins III, Director of the Cyberinfrastructure, National Science Foundation
* Jay Boisseau, Director, Texas Advanced Computing Center
* Jonathan Schwartz, CEO and President, Sun Microsystems
* Hector Ruiz, Chairman and CEO, AMD
Friday June 29, 2007 Sun's switch is bigger than yours
The Register offers their take on the Sun Constellation System:
Sun has teamed with the Texas Advanced Computing Center (TACC) in Austin to build the first "Constellation" system. The box dubbed "Ranger" is a Linux cluster with a peak performance in excess of 420 trillion floating point operations per second. The system is packed with 1.7PB of disk storage, and when finalized, will have 105TB of memory. The setup will take about 3 megawatts of power to run.
The 420 TFlops would put the TACC system ahead of a Blue Gene box at Lawrence Livermore Lab, which has hit 280TFlops. IBM, however, is in the midst of upgrading that system, so it remains to be seen how the Sun kit will compare with IBM when the November Top500 supercomputer list comes out. (A mid-year version of the list will appear this week.) Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 29, 2007 05:00 AM ) Permalink
Thursday June 28, 2007 
I've been working with the people behind the Sun Constellation System this past week in Dresden, including Andy Bechtolsheim. I can't really describe how fascinating it has been to spend time with some of the brightest minds on the planet, but reading this story about Andy reminds me of something:
I was working at Cray Research back in 1996 when Seymour Cray died. He had long left the company to do other ventures, but I was asked to produce a video about the man and his life.
I dug up films from the 70's and watched hours of footage of him speaking and working. And then some interviewer asked him how he wanted to be remembered.
He said he would probably be remembered as just someone who did his work.
And you know what? I think I want the same thing.
And when you read about Andy B. in this NY Times story and wonder why a billionaire would work as hard as this man, ask yourself this:
"He does his work. Is there any better legacy than that for any of us?" Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 28, 2007 02:57 AM ) Permalink | Comments [0]
Tuesday June 26, 2007 Product page up: Sun Constellation System
The good folks over at sun.com/hpc have posted a Sun Constellation System product preview page.
Welcome to Age of Open Petascale Computing! We'll be unveiling the thing at the ISC in just a few hours ;-)
Posted by redbeetle [Sun HPC General Topics] ( June 26, 2007 07:27 AM ) PermalinkSun to Unveil World's First Open Petascale Architecture
I'm writing to you from the International Supercomputing Conference in Dresden, Germany. Today, Sun founder Andy Bechtolsheim will unveil the Sun(tm) Constellation System, the World's First Open Petascale Architecture.
What is it?
The Sun(tm) Constellation System is an open systems Petascale computing environment packaging ultra-dense high performance computing, networking, storage and software into an integrated system that delivers massive scalability, dramatically reduced complexity and breakthrough economics to the HPC market.
What does this mean to Sun in High Performance Computing?
Sun can now offer a complete solution to customers who want to build the world's largest compute clusters. We can compete and win the biggest deals on the planet.
As reported on Marc Hamilton's blog, the first customer installation of the Sun Constellation System is the TACC Ranger cluster. With over 500 TeraFLOPS of CPU power, this cluster will be one of the most powerful general-purpose computing platforms in the world when fully deployed in late 2007.
What's our secret sauce here?
* Ultra-dense IB Switch: The bigger a compute cluster, the more important the interconnect is to efficiency and performance. Until now, the largest InfinBand switch available has had 432 ports. With 3456 ports, the Sun ultra-dense switch helps reduce cost and overcomes the complexity typically associated with large-scale HPC clusters. This single, fully non-blocking monolithic core InfiniBand switch replaces hundreds of discrete switching elements, helping to reduce the number of required leaf switches and vastly simplifying cabling and administration time by up to 80 percent.
* Ultra-dense Blade server platform: With up to 48 blades in a single rack, the Sun Constellation System has incredible compute density with built-in IB connectivity. We're talking 6 TeraFLOPS per rack.
What is Sun previewing in Dresden at the ISC conference?
Andy Bechtolsheim will unveil the Sun Constellation System in the ISC Sun Booth at 6:00 pm CET on June 26. The ultra-dense Switch and Blade Server chassis will be in the booth for customers to see for the first time anywhere. As this is a technology preview, the Systems are static and will not be powered up.
When will this be available for new customers?
The Sun Constellation System is expected to start shipping towards the end of 2007.
For more information including a datasheet and FAQ, please check out the Sun HPC site
I'll be posting photos and video from the show floor very soon.
Cheers,
-Rich Brueckner
editor, Sun HPC Watercooler
Monday June 25, 2007 New Blueprint: Grid Engine Software at Tokyo Institute of Technology
The third segment of a Sun Blueprints series documenting the architecture and deployment of the Tokyo Tech TSUBAME grid is now available online as of today:
Sun N1 Grid Engine Software and the Tokyo Tech Super Computer Grid by Minoru Hamakawa
Abstract
One of the world's leading technical institutes, the Tokyo Institute of Technology (Tokyo Tech) created the fastest supercomputer in Asia, and one of the largest outside of the United States. Using Sun x64 servers and data servers deployed in a grid architecture, Tokyo Tech built a cost-effective, flexible supercomputer that meets the demands of compute- and data-intensive applications. Built in just 35 days, the TSUBAME grid includes hundreds of systems incorporating thousands of processor cores and terabytes of memory, and delivers 47.38 trillion1 floating-point operations per second (TeraFLOPS) of sustained LINPACK benchmark performance and 1.1 petabyte of storage to users running common off-the-shelf applications. Based on the deployment architecture, the grid is expected to reach 100 TeraFLOPS in the future. This Sun BluePrints article provides an overview of the Tokyo Tech grid, named TSUBAME. The third in a series of Sun BluePrints articles on the TSUBAME grid, this document provides an overview of the overall system architecture of the grid, as well as a detailed look at the configuration of the Sun N1 Grid Engine software that makes the grid accessible to users. Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 25, 2007 05:00 AM ) Permalink
Saturday June 23, 2007 Bringing the Grid to Its Full Potential
Grid Today has a story on the EC-GIN, a European-funded project which aims to exploit the network in a way that better suits the needs of grid applications:
Are the network requirements of grid applications different from those of other applications? How is grid traffic different from other types of traffic? What sort of network enhancements do we need in order to improve the performance of grid applications? These and other questions are examined in the work done by EC-GIN, a European-funded project which aims to exploit the network in a way that better suits the needs of grid applications. Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 23, 2007 05:00 AM ) Permalink
Friday June 22, 2007 Taking Licensing to the Next Level
Grid Today editor Derek Harris writes about a new licensing model that could spark an evolution in the way users pay for their grid middleware:
"We have hardware, so we want people to discover that it's useful to run things faster, for example. If there's a huge licensing or a huge economic step to go through before you even try to run it on twice as much hardware -- we don't like that."
Posted by redbeetle [Sun HPC General Topics] ( June 22, 2007 05:00 AM ) Permalink
Thursday June 21, 2007 Wall St. Balancing Performance And Energy Efficiency
Wall Street is no stranger to high-performance computing (HPC). The HPC industry's unofficial tag line once was "Performance at any cost." Today, HPC vendors realize there is a cost associated with speed and that there is a limit to how much firms can pay for increased performance. Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 21, 2007 02:01 PM ) Permalink
Tuesday June 19, 2007 The speed of collaborative research based on Europe's network of supercomputers has become quicker than ever thanks to a new major upgrade. The EU-funded project, Distributed European Infrastructure for Supercomputing Applications (DEISA), has increased its connectivity speeds ten-fold to 10 gigabits per second, through the GÉANT2 pan-European research and education network.
DEISA aims to provide scientific researchers with access to a European cluster of state-of-the-art high-performance computing (HPC) resources. The 'private network' of point-to-point links deployed by GÉANT2 should enable researchers to gain faster and more efficient access to DEISA's shared file system, which in turn supports ground-breaking applications in computational sciences. DEISA's goal is to create an integrated European HPC ecosystem before the end of the decade. Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 19, 2007 05:00 AM ) Permalink
Saturday June 16, 2007 McNealy on the road to bridge digital divide
As reported in The Independent, Sun's Chairman of the Board is out there working the cause as hard as ever:
'Mr. McNealy will have clocked up 40,000 miles on this one-week trip. His thoughts, as his head hits the standard-issue airline pillow, will be turning to the next visits - Europe and later Washington DC - rather than to those halcyon early days, when Sun was a pioneering upstart, built by four friends out of a project at California's Stanford University, and when this multibillion- dollar future lay before it." Full Story
Posted by redbeetle [Sun HPC General Topics] ( June 16, 2007 05:00 AM ) Permalink
Friday June 15, 2007 Japan to build 10 petaflop system by 2009
The Japanese news agency Nikkei reports the Japanese government research institute Riken has appointed Hitachi, NEC and Fujitsu to jointly develop the world's fastest supercomputer by March 2009. Japan's Science and Technology ministry has budgeted 115.4 billion yen ($948.08 million dollars) to build it. The three companies are to design the supercomputer to be able to compute at 10 petaflops, according to the report. The report predicts that the next-generation supercomputer's architecture will be a multi-system combining an accelerator-embedded scalar supercomputer and a vector supercomputer.
Currently, the fastest supercomputer in Asia is the Sun TSUBAME cluster at Tokyo Tech.
Posted by redbeetle [Sun HPC General Topics] ( June 15, 2007 05:00 AM ) Permalink
Monday July 17, 2006 Trade Up from SGI and Save on Sun x64
Today you can trade in your SGI server or workstation and upgrade to a Sun x64 server to enjoy generous allowances. Sun's Upgrade Advantage Program provides the maximum standard trade-in allowance on SGI products upgraded to Sun x64 servers and workstations. Full Story Posted by redbeetle [Sun HPC General Topics] ( July 17, 2006 08:50 AM ) Permalink
Saturday June 10, 2006 60 days on the World's Fastest Workstation: Priceless
In HPC, price/performance is often a key metric. But now, with Sun's "Try and Buy" program, you can "do the math" on some of the fastest workstations on the planet and pay only if you keep the machine past the 60-day trial period. Try and Buy now includes the Sun Ultra 40, Sun Ultra 20, and Sun Fire x2100 servers, all based on AMD Opteron. link Posted by redbeetle [Sun HPC General Topics] ( June 10, 2006 05:00 AM ) Permalink
Tuesday June 06, 2006 Greg was probably the hardest working man I've ever met and a true friend. Rest in Peace, Greg. We'll miss you out there on the road and forever in our hearts. Posted by redbeetle [Sun HPC General Topics] ( June 06, 2006 06:57 PM ) Permalink
Tuesday May 16, 2006 Sun's Chief HPC Evangelist Moves to Head Systems Group
John Fowler, who's been Sun's chief HPC evangelist for the past couple of years, has moved on up to head the newly-formed Systems Group. John's expanded role covering both SPARC and x64 systems is great news for the Sun HPC community. Full Story Posted by redbeetle [Sun HPC General Topics] ( May 16, 2006 10:33 AM ) Permalink
Wednesday May 10, 2006 Trade in Itanium workstations for Sun Ultra 40
Now that winter's over, you probably don't need that Itanium space-heater any more. Why not trade it in for the fastest workstation on the planet?Sun is running a special promotion: Trade-in any qualified Itanium based workstation for an eligible Sun Ultra 40 workstation and get a 20% trade-in allowance. This trade-in allowance is double the normal trade-in allowance but is for a limited time only. link Posted by redbeetle [Sun HPC General Topics] ( May 10, 2006 01:00 AM ) Permalink
Monday May 01, 2006 Special offer: 60-day Free Trial of Sun Fire x64 servers
For a limited time, Sun is offering Free 60-Day product trials on Sun Fire X4100 and X4200 servers.Sun Fire X4100 and X4200 servers, the latest additions to the Sun Fire x64 server family, deliver a new level of industry standard computing for HPC by addressing energy efficiency, reliability and performance issues head on. Powered by multi-core or single-core AMD Opteron processors, these x64 servers run Solaris OS, Linux, and Windows. link Posted by redbeetle [Sun HPC General Topics] ( May 01, 2006 03:46 PM ) Permalink
Thursday March 23, 2006 Sun HPC Cluster Tools 6.0 Software released
Sun HPC ClusterTools 6 software has been released and is available as a free download. With support for Solaris 10 on x64, this software is a key element of the Sun HPC software stack.Sun HPC ClusterTools 6 software is an integrated toolkit that allows developers to create and tune MPI applications that run on clusters and SMPs.Sun HPC ClusterTools software offers a comprehensive set of capabilities for:
* Parallel program development
* Resource management
* System administration
* Cluster administration
Sun HPC ClusterTools 6 software provides developers with the Sun Cluster Runtime Environment (CRE) and the Sun Message Passing Interface (Sun MPI). It is also compatible with Sun N1 Grid Engine, LSF, and PBS distributed resource managers.
Sun HPC ClusterTools 6 software supports:
* SPARC and x64 systems
* Solaris 10 Operating System
* Sun Studio 8,9,10,11
* Shared memory and TCP communication
* The full MPI-2 specification
Product information and Free download Posted by redbeetle [Sun HPC General Topics] ( March 23, 2006 01:06 PM ) Permalink
Thursday March 16, 2006 AMD/Intel performance highlights "compiler factor"
With Intel turning up the hype on how they will be fast as AMD "Real Soon Now," it is encouraging to have third parties illustrate the difference the right compilers make on system performance--especially when the comparison makes Sun look this good! link Posted by redbeetle [Sun HPC General Topics] ( March 16, 2006 08:02 AM ) Permalink
Monday February 20, 2006 Ways to run Dtrace with MPI programs
Sun's Terry Dontje blogs on how to run Dtrace on MPI programs running under Sun’s HPC Cluster Tools environment."Dtrace makes it really easy to trace programs and MPI with over 200 function defined in the MPI standard dtrace provides an easy way of narrowing down what one wants. link Posted by redbeetle [Sun HPC General Topics] ( February 20, 2006 03:22 PM ) Permalink
Registration Open: SunHPC Workshop and Consortium Meeting
Registration is now open for the Combined Sun High Performance Computing Workshop and Consortium Meeting, March 13 - March 17, 2006, at RWTH Aachen University in Germany.The HPC Consortium is a Sun Global Education and Research worldwide customer forum which meets to research and discuss high-end scientific and engineering computing issues and trends. While the Consortium is primarily a technical computing user group meeting for Sun's academic and research customers, customers and interested parties from industry and government are welcome.
The Aachen meeting will particularly focus on developers' interests. Thus we invite participants to profit from the unique opportunity to also take part in the SunHPC workshop in order to learn more about optimization and parallelization and gather hands-on experience. link Posted by redbeetle [Sun HPC General Topics] ( February 20, 2006 03:11 PM ) Permalink
