NASA ditches Itanic for new Xeon-based SGI giant
20,480 cores on the Moon
NASA has once again turned to SGI for a massive supercomputer.
The two organizations announced today that SGI will build a whopping 20,480-core system for NASA Ames in Mountain View, California. The giant will run on four-core versions of Intel's Xeon chip and should reach peak performance of 245 Teraflops, which would make it one of the fastest computers in the world. The system, centered around SGI's Altix ICE hardware, will also boast - gulp - 20,800 GB of memory and 450TB of storage.
The computer will also be water-cooled and run Suse Linux.
NASA plans to use this unit to help with jobs related to future manned missions to the moon and for aeronautics and science research.
This system is of particular note because of SGI's past at NASA. A few years back, SGI built a 10,000 processor supercomputer for NASA around Intel's Itanium chips. That computer - dubbed Columbia - was one of the top three boxes in the world during its heyday.
NASA appears to have shunned the Itanic this time around, which makes us wonder if we'll ever see Itanium crack the top ten again.
In addition, SGI and Intel helped NASA out an awful lot on the Columbia box, offering up hardware at a very steep discount. We're waiting to hear how much this system will cost. ®
Thanks, that makes sense. So it is basically a collection of PCs ganged together on a fast net, not some kind of Illiac-style matrix cruncher.
Gerhard: Windows and Linux both can run on 128 core machines. But that doesn't mean the kernel code will not run 128 times faster, but only 4 or 8 times faster. That's because an OS kernel has a certain number of global resources that have to be interlocked. High performance database systems (and I'm sure SQL Server and Oracle are well ahead of MySQL on this) are designed to minimize resource contentions, and they go out of their way to not invoke the kernel any more than they have to. For example, on Windows you can do asynchronous disk I/O, and often a database system will poll for completion instead of asking the kernel to alert them or wait for an event.
Whiz - bang - thud
...and then Steve Balmer skulks in and loads Server 2k8 on it and performance increases exponentially.......
oh no ...
i just figured out what causes global warming : The computers used to model the weather patterns and check for global warming !!!
Where is Cray when you need them ... A Cray is the only computer known to run an infinite loop in just under 4 hours .. then again SGI used to own Cray...
Mines the one soaked with fluorinert