Posted on 07/21/2008 10:05:05 PM PDT by Swordmaker
Mellanox Technologies, Ltd. today announced that its 40Gb/s InfiniBand technology interconnects the new 29TFlops computer systems research cluster at the Center for High-End Computing Systems (CHECS) within the Virginia Polytechnic Institute and State University. 40Gb/s InfiniBand was introduced in June 2008 and this compute cluster showcases the quick industry acceptance and adoption compared to other high speed interconnect technologies. Mellanox 40Gb/s InfiniBand technology provides the necessary scalability and performance for the CHECS research activities and will be the foundation for the development of next generation, power-aware high-end computing resources.
“Mellanox continues to lead the high-performance computing and enterprise data center industry with the highest performing interconnect technology that delivers unmatched performance for parallel compute clusters,” said Thad Omura, vice president of product marketing at Mellanox Technologies, in the press release. “In 2003 we partnered with Virginia Tech to build the first 10Gb/s InfiniBand large-scale cluster that was ranked number three on the Top500 list at the time. We are excited to partner with Virginia Tech for the first large-scale cluster installation using 40Gb/s InfiniBand, which leverages the mature InfiniBand ecosystem to support the ever increasing demands for efficient and scalable compute systems.”
The new system consists of 324 Apple Mac Pro Servers with the total of 2592 CPU cores efficiently connected with ConnectX 40Gb/s InfiniBand adapters and InfiniScale IV switches, and 40Gb/s copper cables from Amphenol and W.L. Gore & Associates. The system’s performance is ranked among the Top 100 systems of the June 2008 Top500 list of supercomputers. CHECS will be leveraging the system and the new InfiniBand capabilities of congestion control and adaptive routing for expanding high-end computing research activities, in particular in the areas of power aware systems, transparent distributed shared memory systems and high-performance distributed storage systems.
“Our mission is to build computing systems and environments that can efficiently and usably span the scale from department-sized machines to national-scale resources, and will meet the day-to-day needs of computational scientists,” said Srinidhi Varadarajan, Director of CHECS at Virginia Tech, in the press release. “Mellanox’s 40Gb/s InfiniBand technology brings the necessary capabilities to our research activities and will be the focus for the design and deployment of the next generation of high-end systems.”
You, too, could build your own Mac Supercomputer—you can buy the computers for just $906,876 at retail. The previous Virginia Tech supercomputer cluster used quite a few more Mac G5 xServes... but that cost around $5 million.
If you want on or off the Mac Ping List, Freepmail me.
Can’t afford one, after the ritualistic earnings-day massacre.
Do I really need that much for e-mails, Freeping, and a comprehensive climate change model that will finally end the CO2 BS?
why build a cluster out of guchi machines ?
A bunch of NVidia Graphic Processors working in vanilla PCs with low tech CPUs for a fraction of the money of this stylis cluster would have offered a lot more flops for the buck.
And no one has done it because ...?
*Drools*
So many macs.
Graphics processors would not do the heavy lifting that 2,592 3.2GHz Intel® Xeon® processor cores with dual 1.6GHz FSBs with access to 32 GBs of buffered, error correcting RAM can accomplish. This cluster probably needs only a couple of Graphic processors to drive video outputs of the primary controlling nodes.
Why use "guchi" Macs to build this? Sinple: it is actually less expensive. To duplicate the specs of a Mac Pro, a Dell with matching Xeon processors would cost more than $1,000 more than the off the shelf Mac Pro.
Maybe because they’re all defective, overheat and die? :P
In the 70s, we used string to hundreds of TI amplifier chips together to build +1KW radio transmitters. There always came a point when fires were inevitable.
VT has resold the machines in their clusters. The high resale value can really reduce the overall cost.
Here’s an example...
http://www.macmall.com/macmall/promotions/custom.asp?p=supercomputer
Ahem... what’s the carbon/electronic footprint of this computing beast?
Warning:
In the future, the government will begin taxing you on your “electro-carbon” footprint - that is, how many electrons you consume each month, presuming that the higher consumption of them results in a greater generation of carbon from the fuel needed to generate those electrons.
Wait and see!
- John
Because Mac Pros are pretty cheap compared to the competition. Plus Macs come with XGrid, which makes supercomputing very easy.
A bunch of NVidia Graphic Processors working in vanilla PCs with low tech CPUs for a fraction of the money of this stylis cluster would have offered a lot more flops for the buck.
GPUs are very fast for the types of calculations they can perform. Unfortunately, those types of calculations are very limited.
I tried using it, but when I moved the mouse just a little, the cursor pointer flew off so hard it smashed right through the side of the monitor.
When I decided to use it, it was finished three hours before I started. Now that's fast!
I walked by it a minute ago and it looks like the old one.
Xgrid, for those not in the know, is a suite of software, included with the Mac OS that enables any Mac - ANY mac - on a network to participate in distributed computing on a very large scale. Seriously cool software that works right out of the box...
Even my $23 shareware video conversion software, VisualHub, uses XGrid to crunch videos across all the Macs on your network. Now I just need another Mac...
Tachyon computing — the next big thing...
The Folding@Home project is using NVidia GPUs to help fold proteins, and for this the GPU is a perfect solution. It is able to simulate the thousands of atoms in a protein incredibly fast.
FWIW, a single NVidia G280 would generate more points (completions) for the owner than any other FReeper Folder and would place them as #1 on our team, at about 7,000 ppd.
There are some limitations to its calculations but what the NVidia crunches it crunches very, very fast.
They still have the re-made supercomputer from 2004, the one made from 1100 dual processor G5 xServes. That one has 2200 processors. This is a new one that has only 324 boxes... but in each of those boxes are 8 processors (cores), making a total of 2,592 processors. Apparently it is not in the same location as the other one.
I can’t imagine another building with the infrastructure to host it unless it is in the VBI building.
It probably is in the original building. The original Mac supercomputer was made with 1100 full size Mac Towers. A year later they replaced those with 1100 Mac xServes which are rack mounted and take up 1/3 the volume of the Towers. That means the space required was 1/3 less. They could install the 324 MacPros in half the space that was freed in the rebuild of the original machine.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.