While there’s isn’t exactly a shortage of cloud providers offering arrays of CPUs that wouldn’t be out of place in a science fiction movie, there is nowhere near the same choice when it comes to GPUs (Graphics Processing Units). This is changing, however. Last year Peer 1 Hosting, a global online IT hosting provider, delivered what it described as the industry’s first large-scale, hosted GPU cloud. Amazon Web Services also started offering cloud-based GPU resources in November 2010.
Both cloud providers currently use Nvidia hardware. Peer 1 Hosting runs the RealityServer 3D web application service platform, developed by Nvidia-owned mental images. This uses Nvidia Tesla GPUs and 3D web services to deliver interactive and photorealistic applications over the web using the iRay renderer.
Of course, demand for GPU-based cloud computing will come from the software sector and last month Bunkspeed, the developer of specialist 3D rendering and animation software, introduced its Bunkspeed Cloud Solution. It utilises Bunkspeed Shot Pro and mental images’ RealityServer and helps users publish interactive photorealistic scenes or 3D content to the Web. These can then be accessed from anywhere in the world, on almost any device via a web browser. It also helps non-CAD decision-makers view and review such projects.
Bunkspeed is the first major rendering software developer to fully utilise GPU-based cloud resources and introduce new rendering workflows. In theory, any software developer that uses mental images iRay as a rendering solution should also be able to do this without too much trouble. This includes Dassault Systèmes with Catia Live Rendering, Real Time Technology with RTT DeltaGen and Autodesk with 3ds Max. With Autodesk already showing interest in cloud-based rendering and collaboration, we would be surprised if the company didn’t unveil a GPU-based technology soon.
Of course the question arises: does it really matter to the end user what hardware is being used to render a scene? The answer is: probably not. What are important are speed, cost, quality and the ability to benefit from new workflows, and the market will ultimately decide how important cloud-based GPU rendering will become.