Today, Peer 1 announced GPU Cloud powered by nVidia Tesla graphic processor units. This means you can rent time on a high-powered graphic server by the hour, day, or month. Last Friday, they expanded their offering to include hosted graphics processing unit Cloud servers located in the UK.
 
Peer 1′s new service offers large-scale computing on demand for applications such as graphics rendering, complex quantitative processing, video compression, and large-model 3D web services.
 

nVidia Tesla M2050 - Fermi GPGPU silicon for HPC Center integration - this time, with Peer1

nVidia Tesla M2050 – Fermi GPGPU silicon for HPC Center integration – this time, with Peer1

 
Instead of expending precious capital and purchasing expensive hardware, you would have to spend $4,700 on a pair of Tesla M2050 [Fermi architecture-based] GPUs which provide up to 1 TeraFLOPS of double-precision performance. Then you have to add the cost of a 1U Supermicro with dual-quad core Intel-based server which can easily add additional $2,000. Peer 1 can rent time accessible, via the Cloud, at $2.50 per hour per GPU or $40 per day per GPU.
 
Peer 1 is also offering the RealityServer 3D web application software developed by Mental Images on the platform. RealityServer delivers interactive, photo-realistic applications over the web using the iray renderer, allowing animators, product designers, architects and consumers to easily visualize 3D scenes with remarkable realism. By hosting the GPUs and software within the cloud this simplifies customer implementation and provides an inexpensive entry point for new users. This leaves your staff free to focus their talents on developing new projects.
 
Sumit Gupta, product line manager of Tesla products at nVidia, said "With Peer 1′s commitment to best-in-breed technology and state-of-the-art data centers, the Cloud Hosting has all the skills to deliver a service of this kind and make the power of the GPU more accessible than ever before."
 
This BSN* link gives more information about the Supermicro 1U GPGPU oriented server line. Watch for more articles from the nVidia GPU Technology Conference 2010, which officially starts tomorrow.