NVIDIA-based cloud service offers GPUs for rent

Print Friendly, PDF & Email

PEER1 Hosting announced from SIGGRAPH yesterday in Los Angeles that their GPU-powered public rendering cloud is up and going. From the press release

nVidia logoThe system is running the RealityServer 3D web application service platform, developed by mental images, a wholly owned subsidiary of NVIDIA. The RealityServer platform is a powerful combination of NVIDIA Tesla GPUs and 3D web services software that delivers interactive, photorealistic applications over the web using the iray renderer, enabling animators, product designers, architects and consumers to easily visualize 3D scenes with remarkable realism.

With the use of massively parallel NVIDIA Tesla GPUs PEER 1 Hosting can now offer customers flexible and reliable access to a system capable of delivering high computational performance across demanding applications such as graphics rendering, complex quantitative processing, video compression and large-model 3D web services for access by mobile clients.

We’ve talked about these technologies before. According to NVIDIA, “more than 128” Tesla S1070 and Tesla M2050 (Fermi architecture-based) cards are in the system (the exact number has not been disclosed) GPUs as well as RealityServer are now available for purchase worldwide and will be hosted as a managed hosting offering at PEER 1 data centers in Toronto, Canada and London, UK.

According to PEER1, pricing is structured on a per GPU per month basis for starters. A server with 4x s1070 GPU and 2x Intel processors starts at around $2000 a month fully managed.  Discounts are available for long term contracts and premiums are applied if you need less than one month. That isn’t cheap, so only users with a business model (or research funding) need apply.

Why mix the Tesla 10- and 20-series? Some of PEER1’s customers may want the 4GB memory of S1070 GPUs vs the 3GB of M2050. Also, Peer1 bought their S1070s before the M2050s became available.

Comments

  1. On the face of it, this sounds like a good idea and is part of the evolution of cloud for the HPC environment. Firstly, GPUs are good, proven performers in the computation stakes. See my other posts on GPUs. Second, using applications like RealityServer makes sense because it enables customers to interact with their data (or specifically models, 3D content in this situation) over the web.

    However, for those customers not using web based applications like RealityServer, use of the cloud in the HPC environment can pose a problem. Firstly, do customers have the bandwidth available to take receipt of the data (which can be multiple Terabytes or Petabytes once processed)? Second, does the cloud service provider, which could have hundreds (potentially) of customers each generating multiple Terabytes of data, have the bandwidth available to send the data back to their customers? Finally, does the customer have the appropriate infrastructure to store the data created during the process?

    These problems can be overcome – certainly integrators like OCF can provide adequate storage facilities for customers – but none the less, customers must consider these questions before proceeding with a cloud service for HPC related problems.