GPU rendering – Is it worth it?

As you probably know, GPU rendering harnesses the power of a graphics processing unit for rendering purposes instead of relying on CPU (a computer processor). Using GPU for rendering is quickly growing in popularity across the world for a variety of use cases, from animation and VFX projects through to architecture and interior design applications, but is GPU rendering worth it?

When you rely on a GPU for rendering, the most important element of the GPU is the amount of memory available. The more memory a GPU has, the more powerful and efficient it can be. A lot of people don’t realise that memory is even more important than the overall speed of a GPU in determining the speed of a render, but a fast GPU with less memory will often lead to renders frustratingly crashing.
In an article on cgsociety.org, Panos Zompolas, the founder and CTO of Redshift Rendering Technologies, said: “An individual artist can install four high-end GPUs onto a single computer and rival the performance of several tens of high-end CPU-based render farm machines. And you get all of this with lower electricity costs and a fraction of the occupied physical space compared to what an equivalent CPU render farm would require.”

He’s right of course. According to some industry tests, it is more than 15 times faster to render on a single GPU than a CPU processor, but speed isn’t everything and while using GPU has been touted as a rendering solution for the future, the industry is currently waiting for software vendors to make more GPU enables rendering platforms.

It is true that GPUs paired with technology developed by Redshift can render images significantly faster compared to CPU renderers, and it sometimes follows that faster rendering does mean lower hardware costs, especially considering that a single GPU can equal the rendering performance of five-to-twenty high-end CPUs on average.

However, many businesses already have on-premises render farms which they have invested in and therefore a hybrid model which incorporates: GPU rendering, existing CPU infrastructure and cloud rendering services (which are scalable on-demand), is better positioned to save a business money in the long term, while still enabling it to take full advantage of the efficiency and control which GPU rendering offers.