NVIDIA introduced the expensive, but "sticky" Tesla supercomputer

AMD has already announced plans to create a platform for clustered computers, and therefore add fresh gigabytes megahertz processor, but NVIDIA overtook an opponent on the floor of the case with yesterday's release of the GPU- based platform Tesla Personal Supercomputer.

This very supercomputer promises to deliver the impressions of a real cluster in only 1/100 of the price - in my opinion it is very profitable, the only catch is that, in fact, it is completely incomprehensible how much to pay. Each platform is sold in a separate office, which sometimes works exclusively by telephone. I'll tell you the details: there are 8 platforms in total (as far as I understood in the search process for deep American resources, all processors: AMD), the difference in their 2-3-4 GPU components. Each of them works on the basis of computing NVIDIA Tesla C1060 GPU which, in turn, is built according to the architecture of NVIDIA CUDA - it is a technology for parallelizing calculations and therefore “provides the user” with 960 cores.

Such computers will be produced on the basis of the new platform by both ASUS, and Dell and Lenovo (and even several lesser-known brands). For this reason, one of the coolest household monsters can cost $ 8,495, but I collected something more interesting for myself, albeit more expensive: $ 17,000. The result of an adventure under the cut.

Specifications station BOXX Personal Supercomputer.

• Processor: AMD Phenom 9950 Processor (2.6 GHz) (Quad-Core)
• RAM: 4x 4Gb DDR2 800 (DDR3 is available only in a single science-oriented station)
• Primary Drive: 300GB 10,000rpm SATA
• Additional drives: 4 x 750GB 7,200rpm SATA (total: 3Tb)
• Drives: 2 x 20X Dual-Layer DVD + -RW Writer
• Computing Processors: 4 x Nvidia Tesla C1060 PCIE
• Operating system: SUSE SLES10: Upgrade and patches, individual technical support - 3 years
• Standard black keyboard
• Mouse: Logitech MX Revolution

Rummage and try to collect "your cluster" fans really worth it - there are a lot of interesting things, but it would be more pleasant to work for the animal live. I think the option in accordance with the current crisis can be thought up for $ 6-7,000, but who saves on supercomputers?