A Quick Look At OctaneRender 2019 RTX Performance

It’s easy to love OTOY. The company not only produces one of the most feature-rich and capable renderers available, OctaneRender, it also develops a relevant benchmark around it, OctaneBench. We’ve been testing with OTOY’s benchmark for a while, most recently in our look at NVIDIA’s Quadro RTX 4000. With the upcoming OctaneRender 2019 adding support for NVIDIA RTX, OctaneBench results are becoming much more interesting.

OctaneRender 2019 is bringing a new rendering engine called Brigade, and with it, OTOY is able to take advantage of the latest GPU technologies, including RTX, DXR, and Vulkan. The use of Vulkan means that this future Octane release will add support for AMD’s Radeon GPUs, which suffice to say, is a great thing.

We talked to OTOY at NVIDIA’s GTC this month, and were told that once OctaneRender 2019 ships, the use of RTX’s accelerated features will be suitable for final frame rendering. Ultimately, this is what we’d like to see from any company developing a renderer around accelerators, such as the RT cores. As far as we understand it, every renderer promising RTX support should be suitable for final frame rendering. That includes V-Ray, Redshift, Arnold GPU, Dimension CC, RenderMan, and SolidWorks, to mention a few.

At the moment, the primary version of OctaneBench is 4.00, which supports current GPUs, but not the special processors on RTX cards. To test those right now, you need a special version of OctaneBench, conveniently found on the official forums. This benchmark is considered a preview release, but it does a great job in highlighting the kind of speed-ups people can expect when utilizing RTX.

To generate some quick results, we tossed every RTX card we have into the workstation rig, and let them rip. Here’s a quick look at the test PC used, and then the results:

It’s important to note that these scores are composed of three separate scores, representing performance for three different kernels: info channels, path tracing, and direct lighting. Info channels sees around a ~4.5x performance speed-up with “RTX On”, but it’s the least important of them all. Path tracing and direct lighting account for 90% of the overall score, and both see around a 2.5x speed-up with RTX on.

The scaling seen across these five cards isn’t too interesting, only because it scales just as we’d expect it to. The bigger the GPU, the better the performance. In the 2080 Ti’s case, it’s about 33% faster than the 2080, both with RTX on and off.

What these scores ultimately amount to is quicker render times for the faster cards. A 33% advantage may not seem too significant when we’re just talking about a score, but higher scores here means greatly improved performance – hugely important in a real-time rendering engine like Octane.

OctaneRender 2019 isn’t out quite yet, and this surely won’t be the only time you’ll be hearing about this renderer from us. Once the final build of OctaneBench RTX is released, we’ll retest and see how things have changed. As it stands now though, it’s not hard to be impressed by the massive speed-ups seen with RTX turned on here, and given we’re dealing with final frame renders, the results are relevant to absolutely every OctaneRender user.