Traditionally, 3D rendering and video encoding have been extremely processor intensive, often leaving artists twiddling their thumbs for many hours at a time while they wait for their workstations to render their final projects. Then came GPU acceleration.
Although this buzzword’s been around for a couple of years now, I believe the technology is finally starting to come of age, as it is being integrated into industry-standard tools such as 3ds Max and Adobe’s design suites. This means artists and designers can take advantage of the technology without having to alter their workflow to include new and unfamiliar tools or software packages.
For 3ds Max…
Autodesk’s 3ds Max and 3ds Max Design 2012 are both on the GPU acceleration bandwagon. They now ship with iray, a GPU-accelerated renderer based on mental ray. You also have the option of using them with V-Ray 2.0. Both these render engines are physically accurate, making them ideal for the kind of photorealistic work that’s called for in architecture and design visualisation. They can render upwards of ten times quicker than their CPU-based counterparts, making it possible to add materials and light a scene with realtime feedback at near-final output quality. The interactive quality that I have seen is certainly good enough to base creative decisions on, and the final renders take minutes instead of hours. (As an added bonus, they both have pretty minimal setup times.)
In Creative Suite…
Adobe has also been working hard to add GPU acceleration to many of their packages, with Flash and parts of Photoshop now being accelerated on the GPU. They have also added GPU accelerated features to After Effects and Premiere. These programs now offload playback onto the GPU, making it possible to play back more streams of higher resolution video while you edit.
Certain effects are now also accelerated on the GPU, increasing the efficiency with which they’re applied and rendered, and reducing the load on the CPU so that it can concentrate on other tasks. This means that artists can play back their timelines and render their effects much more smoothly and see a broadcast-quality version of their work far sooner, even with today’s demands for ever higher resolution and deeper compositions. There’s far less time spent waiting for your machine to catch up with your imagination.
Now that GPU acceleration is built into these industry-standard tools, it has made it easy for content creators to take advantage of this exciting technology. Aside from their support for the most popular languages for creating GPU-accelerated applications (CUDA and OpenCL) and support from many software vendors, NVIDIA’s Quadro cards are also available with up to 6GB of RAM on board, allowing them to keep much of their processing local to the card rather than slowing things down by swapping data to and from system memory. They’re a great choice and well worth a look if you’re thinking of moving to a GPU-accelerated workflow.