Caustic Unveils First Real Time Global Illumination Video

[Jeff Mottle]

The guys over at Caustic graphics sent me a recent demo of their Caustic Graphics Card being used in Brazil/3ds Max

 

Demo Scene Specs:

 

·Shows 5-bounce GI enabling photorealistic effects like color bleeding

·Full interaction with all geometry, including high-polly-count models Interactive geometry deformation, not just translation Interactive material editing

·Many different materials interacting in the scene

 

About the Demo

This is our latest interactive demo of "true" global illumination using the CausticRT platform with Brazil in 3dsMax. Its a fully ray traced interior living room scene with 2,080,957 polygons, and outdoor and indoor lighting that includes classic 3D geometry such as the Stanford Dragon & Bunny. The poly count for each of the objects are as follows:

 

Bunny: 70K polygons

Dragon: 800K polygons

Buddha: 1M polygons

Interior: 22K polygons

 

The 3dsMax ActiveShade window resolutions are: 400x300 & 800x600

 

Why is this significant?

 

No one has ever shown such a complex demonstration of true interactive GI using a GPU or CPU implementation. Why? Because GPUs and CPUs are not architected to fully solve the ray tracing problem efficiently. The CausticRT platform addresses these deficiencies and enables GPUs and CPUs to shade with an efficiency comparable to rasterization.

 

The raytracing problem is very different from rasterization. We have seen some impressive demos. Researchers have achieved nearly 100 million rays per second [1]. But there is always a catch. Sometimes the demos show effects that don't really require ray tracing. Often, every object is shaded with the same material. Usually, the scenes are much too small to be representative of real-world use cases.

 

Why haven't we seen a complete ray tracing renderer running on a GPU?

 

To answer that question, we need to take a closer look at the differences between the ray tracing and rasterization algorithms.

 

Rasterization is the perfect example of a streaming algorithm. The scene geometry flows sequentially through a pipeline where each polygon in a mesh can be decomposed into fragments that all need to be shaded with the same shader. This allows the GPU to shade many pixels together, running the shader instances in lock-step on what amounts to a wide SIMD machine. Additionally, because of the locality of the fragments in scene space, it is likely that each instance of the shader will load the same assets and sample nearby texels, thereby getting excellent utilization of a modest size cache.

 

Unlike rasterization, ray tracing allows objects to interact with each other. Inter-object visibility is the key reason that ray tracing enables the stunning visual effects viewers have come to expect from production quality 3D content. But the technical upshot of rays bouncing all over the scene is that a ray tracer needs random access to everything. This means the entire scene (all geometry, shaders, and assets) must be accessible in RAM.

 

Indirect visibility creates another problem for a stream processor when a surface shader uses secondary rays to compute the incident lighting. If two adjacent pixels being shaded in together lock-step emit secondary rays that encounter different materials, those shaders can no longer run together. Additionally, assets needed by the one shader will compete for cache space with assets needed by the the other shader.

[abowen]

great board . reminds me of the PURE cards a 'few' years ago.

can we expect integration with other render engines like VRAY ?

also , will this card handle output render (passes) etc or is it just viewport driven.

it looks like they have addressed what i found most problematic about pure / renderdrives, which is their integration with the renderer.

 

andre

[Jeff Mottle]

great board . reminds me of the PURE cards a 'few' years ago.

can we expect integration with other render engines like VRAY ?

also , will this card handle output render (passes) etc or is it just viewport driven.

it looks like they have addressed what i found most problematic about pure / renderdrives, which is their integration with the renderer.

 

andre

 

Based on Chaos Group's work with CUDA and the speed they have been able to get on their own tests with the GPU version of V-Ray RT, I'd say it's unlikely. Only time will tell though.

[Devin Johnston]

Before VrayRT came out when Caustics first made it's appearance I asked Vlado if they were going to support the card, he said they were going to look into it and that was the last I heard about it. From the what I've seen and heard about the GPU version of VrayRT it sounds like they are getting performance speeds that are comparable to the Caustic card. If that's true then I don't see the Caustics card being supported at all since you will probably be able to buy several moderately priced video cards for what one Caustics card will cost.

 

Jeff, I know you are working with several video card manufacturers right now to analyze their cards for real world performance. I hope when the Caustics card & Vray RT (GPU version) comes out you will be able to do some kind of comparison between them.

[Jeff Mottle]

Jeff, I know you are working with several video card manufacturers right now to analyze their cards for real world performance. I hope when the Caustics card & Vray RT (GPU version) comes out you will be able to do some kind of comparison between them.

[

 

Sure will do. I suspect that will be 12-18 months from now. Caustic should be available in the spring, but VR-RT I suspect will be a bit longer based on track record