Here is a collection of highlights, selected totally subjectively, from this week’s HPC news stream as reported at insideHPC.com and HPCwire.

10 words and a link

Cray forms Indian subsidiary

120 laid off at SGI

Anandtech.com compares Instanbul and Nehalem

Cilk++ parallel development tool launches, compared to OpenMP

Profiling MPI applications using Sun tools

Article: Express Parallelism, Don’t Manage It

OSC boosts bioscience with super expansion

Acceleware’s new GPU services group

Article explores answers for challenges faced by virtualization in HPC

Russian HPC company intros cell-based super at CeBIT

Rocks-A-Palooza IV

TeraGrid conference looking for education, workforce development submissions

Japan’s Earth Simulator 2 open for business

The Mainichi Daily News, the English-language news site of Japan’s Mainichi Newspapers, reported this week that the new earth simulator is ready to go:

Japan’s Earth Simulator 2 (ES2) supercomputer has been unveiled to the public, following a set of upgrades that have more than tripled the speed of the original Earth Simulator.

…
The old Earth Simulator consisted of 640 linked nodes, but newer technology has whittled that figure down to 160. ES2 has a theoretical top computing capacity of 131 teraflops, or 131 trillion calculations per second — 3.2 times more than its predecessor. It consumes 20 to 30 percent less electricity, and the computer’s floor space has been halved to 650 square meters.

Despite the TOP500 heritage of its older brother (the first Earth Simulator spent two and a half years at the number one spot), that peak capability puts it far away from the number one spot on the November 2008 list. In fact, the machine isn’t even the fastest in Asia anymore. Things change.

Ferrari driving more than sports cars

Ferrari is driving innovation in more than just its high performance sports cars these days. Its design processes are so dependent upon high performance computing that they’re beginning to drive datacenter technology as well. It has actually separated its corporate design/IT datacenter from its Formula-1 design center. The latter, aptly named the F1 Data Center, houses roughly sixty racks of IBM, Sun and HP servers and storage. The building in which it is housed was constructed back during the 1940s, so Ferrari hired APC to come in and modernize everything based on its latest datacenter requirements.

“This is not exactly what you call ‘green computing’,” [Piergiorgio] Grossi [CIO of Ferrari F1] said with a smile. “High-performance computing, of course, is a different animal. But we still make sure that we are as efficient as possible in the way we use computing.”

The center services CFD simulations and other technical needs 24×7. During race season, the Ferrari race team actually brings part of the datacenter with them. 150 Ferrari IT members and seven trucks worth of servers and power supplies accompany the race team to 18 locations during the season. That is impressive!

For more info on Ferrari’s F1 datacenter, read the full article at eWeek.

NCAR bill signed by Governor

Governor Dave Freudenthal of Wyoming has signed a bill into law that removes a potential bureaucratic bump in building the new National Center for Atmospheric Research [NCAR] supercomputing center in Cheyenne. NCAR has partnered with the University of Wyoming, the state of Wyoming and the University of Colorado at Boulder to spearhead the new center for climate and atmospheric research.

According to the article, the center is slated to open in late 2010 or early 2011 at a cost of $60 million. Considering the recent NSF Track awards, $60 million is a real bargain for an operational supercomputing center. Wyoming is one step closer to breaking ground. You can read the full article here.

—–

John West is part of the team that summarizes the headlines in HPC news every day at insideHPC.com. You can contact him at [email protected].