March 26, 2009
10 words and a link
University of Manitoba announces HPC facility
Nagoya University goes with Fujitsu on 60 TF super
Red Bull teams with Platform for 2009 racing season
ScalableInformatics intros GPU+Cell box
ANSYS expands HPC capacity for R&D
ACM blogs on improving CS education
My supercomputer lied to me!
Sun sells two new Constellation systems in Australia
insideHPC's man down under sent us a note today letting us know that winner in the competition for new supers at the Australian Bureau of Meteorology and the Australian National University has been announced. From Australian IT:
[The agencies] will spend $30 million on two new supercomputers that will more accurately predict cyclones, tsunamis and the effects of climate change.
The supercomputer, which is being supplied by Sun Microsystems and will be the most powerful in the southern hemisphere, will be ten times more powerful than the Bureau's current system. It will be able to crunch 1.5 trillion calculations per second when it is at full operational capacity in two years time, Bureau of Meteorology chief information officer, Phil Tannenbaum said.
The new supers will be part of an effort to assess both weather and climate change at the same time:
"This supercomputer will allow us to map the effects of climate change with what we call an 'earth model' which can measure both weather and climate at the same time. Previously we were only able to do weather forecasting or climate forecasting, but not both at the same time," Mr Tannenbaum said.
This new 'earth model', or The Australian Community Climate and Earth-System Simulator (ACCESS) as it's known to the scientific community, couples climate and weather forecasting with an earth system simulator. It is being developed as a joint initiative between the Bureau of Meteorology and CSIRO in cooperation with the university community in Australia.
The direction of this purchase had been hotly speculated, since insiders tell me that the Australian Bureau of Meteorology tends to follow UK MET who recently purchased an IBM (and ran into some trouble over the machine's carbon footprint). insideHPC posted an Inside Track back in November that the acquisition had been delayed due to rumored extended negotiations with Sun. Evidently that rumor was right on the money. This purchase displaces incumbents NEC and SGI.
Rackable ready to run hot, innovates on power distribution
Rackable has announced a new scale out computing solution that runs at temps up to 104 degrees F, and moves to per-cabinet (instead of per-server or per-blade) power supplies. The CloudRack 2 refines the CloudRack design announced back in October of last year, which was basically a horizontally mounted cookie sheet with a server on it.
The CloudRack 2 servers don't have fans (those are moved to the back of the rack itself), and they don't have power supplies. From a story at internetnews.com:
At the same time, it's taking out the heat-generating power supply that connects to every rack. Instead, the racks connect directly to DC power plugs. So instead of a power supply for each rack, the cabinet itself has one very large AC to DC converter, and all of the racks connect to that converter.
Without the power supplies, the new servers are rated to run at up to 104 degrees F (40 degrees C), and the extra room gets used to put three servers on a train instead of the two found on the original CloudRack:
The result is a cabinet with 99 percent power efficiency because the DC power draw at each rack is exactly what the rack needs at that time. The C2 cabinets can hold up to 1,280 cores, or 320 processors. Rackable offers AMD Opteron and Intel Xeon processors, and plans to offer the forthcoming Nehalem-EP X55x0 line of Xeons once Intel releases them.
Timothy Prickett Morgan also has coverage at The Register.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?