Endless Paradigm

Quote:Researchers at the University of Antwerp in Belgium have created a new supercomputer with standard gaming hardware. The system uses four NVIDIA GeForce 9800 GX2 graphics cards, it costs less than 4000EUR to build and thanks to NVIDIA's CUDA technology it delivers roughly the same performance as a supercomputer cluster consisting of hundreds of PCs!

This new system is used by the ASTRA research group, part of the Vision Lab of the University of Antwerp, to develop new computational methods for tomography. The guys explain the eight NVIDIA GPUs deliver the same performance for their work as more than 300 Intel Core 2 Duo 2.4GHz processors. On a normal desktop PC their tomography tasks would take several weeks but on this NVIDIA-based supercomputer it only takes a couple of hours. The NVIDIA graphics cards do the job very efficiently and consume a lot less power than a supercomputer cluster.

Here's a look at the specifications of the FASTRA desktop superPC. The main reason why they configured an AMD system is because they couldn't find a motherboard for the Intel platform that could fit four GeForce 9800 GX2 graphics cards. Another interesting note is that this system doesn't need SLI, their application uses the NVIDIA CUDA programming model which makes all eight GPUs work in parallel. The researchers say they don't need SLI during a reconstruction as every graphics card communicates directly with the CPU, no inter-GPU communication is needed.

  # AMD Phenom 9850 processor + Scythe Infinity CPU cooler
  # 4x MSI 9800GX2 graphics card
  # 4x 2GB Corsair Twinx DDR2 PC6400 memory
  # MSI K9A2 Platinum motherboard
  # Samsung Spinpoint F1 750GB HDD
  # ThermalTake Toughpower 1500W Modular PSU
  # Lian-Li PC-P80 Armorsuit case
  # Windows XP 64-bit

The medical researchers ran some benchmarks and found that in some cases their 4000EUR desktop superPC outperforms CalcUA, a 256-node supercomputer with dual AMD Opteron 250 2.4GHz chips that cost the University of Antwerp 3.5 million euro in March 2005:


Basically the ASTRA researchers have created a very affordable computer system that can perform their tomography work very efficiently. For specific applications GPUs are much faster than CPUs, for the tomography work the NVIDIA-based FASTRA system outperform a real supercomputer that's much more expensive, takes up more room and uses a lot more power. The Vision Lab is now planning to build a cluster of such systems, which will allow for real-time reconstruction of large 3D volumes.

That's pretty impressive, but CAN IT PLAY DOOM?


EDIT: SOURCE

Wow, i wish i had that, and only $6,216.01.

Quote:The guys explain the eight NVIDIA GPUs deliver the same performance for their work as more than 300 Intel Core 2 Duo 2.4GHz processors

holy........ poo poo.........

Less than €4000? That's fudgeing amazing.

Quote:The guys explain the eight NVIDIA GPUs deliver the same performance for their work as more than 300 Intel Core 2 Duo 2.4GHz processors

It'll only be this effective for a very small range of uses, for which GPUs are more efficient than CPUs.

Assassinator Wrote:

Quote:The guys explain the eight NVIDIA GPUs deliver the same performance for their work as more than 300 Intel Core 2 Duo 2.4GHz processors

It'll only be this effective for a very small range of uses, for which GPUs are more efficient than CPUs.

yeah shame really....... LIIIIIIIIINUX!!!!!!

quite cool i would like to play around with one of those lol

Assassinator Wrote:

Quote:The guys explain the eight NVIDIA GPUs deliver the same performance for their work as more than 300 Intel Core 2 Duo 2.4GHz processors

It'll only be this effective for a very small range of uses, for which GPUs are more efficient than CPUs.

With the introduction of pixel shaders, that's somewhat different.

GPUs have always been far more powerful than CPUs (If I recall correctly, the top range GPU is about 30-40 times more powerful than a top range CPU) - the main issue is that GPUs are heavily multi-threaded, whereas CPUs are designed to be single threaded (and also, you have to trick the GPU into thinking the data is pixel data).

But since supercomputers primarily rely on multithreaded processing, it's not an issue for them :P


There have actually been a number of non graphics related programs which make use of the vast power of the GPU to do stuff.  I heard that nVidia even had an SDK to help people do it, but haven't really heard much about it since.