Introducing the Ray Tracing Levels System – and what it will mean for gaming

Share on linkedin
Share on twitter
Share on facebook
Share on reddit
Share on digg
Share on email

If you’ve even got one eye on the graphics market you can’t help but have noticed that there’s a considerable buzz around ray tracing – a new paradigm for rendering light in a 3D scene that offers benefits for both gamers and developers. For gamers, it promises much greater realism and immersion by using light in a much more natural-looking way and is a key step towards creating truly photorealistic games, while for developers it offers a simpler and more efficient way of creating lighting. This is because effects such as reflections and shadows, which are challenging to simulate using traditional rasterisation, appear naturally as a result of the ray tracing algorithm.

Of course, while we refer to ray tracing as new, as a concept for 3D graphics it dates back to 1968, with further developments along the way, but due to the computational complexity required it only became practical to implement in real-time relatively recently. Second-generation ray tracing-capable graphics cards for Windows-based desktop PCs have just launched, promising 4K visuals with ray-traced effects for the first time, while the next-generation gaming consoles arriving in the market by the end of the year are also promising ray tracing capabilities. Ray tracing is causing such a buzz that even current generation consoles are getting in on the act, with the classic PC title Crysis being remastered with ray traced effects that will run even on older graphics cards and current generation consoles that do not have ray tracing specific hardware support.

With all this going on, many do not appreciate that not all hardware marketed as ray tracing enabled is created equal. Imagination Technologies has been leading in the ray tracing space for several years and had a development board with working hardware in 2016 that was more sophisticated than the solutions in the market today.

To help the industry understand the differences between the various solutions on offer, we have created the Ray Tracing Levels System – ranging from Level 0 through to Level 5. This consists of:

  • Level 0 – Legacy Solutions
  • Level 1 – Software on Traditional GPUs
  • Level 2 – Ray/Box and Ray/Tri Testers in Hardware
  • Level 3 – Bounding Volume Hierarchy (BVH) Processing in Hardware
  • Level 4 – BVH Processing with Coherency Sorting in Hardware
  • Level 5 – Coherent BVH Processing with Scene Hierarchy Generator in Hardware

The key point that comes across when reading the detail is that as we move up each level there is an increased focus on efficiency. The market however does not always move based on efficiency. For example, the recently released second-generation ray tracing-capable desktop PC hardware has increased performance by simply providing more ray tracing processing – which naturally requires more silicon area and more power. However, this brute force approach is not a viable option for solutions that need to operate within a power budget, such as mobile.

ray tracing Level 4
Images such as these, featuring realistic, dynamic shadows, at 60fps, will be possible in real-time on Level 4 ray tracing hardware.

On console, we can also see that ray tracing effects will only be used sparingly as the hardware does not have the efficiency required for many complex effects. For example, in a demo sequence from one launch title, we can see an object casting ray traced reflections on the floor. However, as soon as that object is destroyed the reflections disappear in an instant. This is because it is not possible to update the scene geometry fast enough, as this would require a scene hierarchy generator in hardware – which we describe as a Level 5 solution. Equally, some early titles we can see show evidence of single bounce reflections and low-resolution reflections, as without a coherency gathering solution (as per our Level 4 solution), these effects are again too computationally expensive or execute with poor efficiency.

Gaining an understanding of ray tracing architecture is essential to put the ever-increasing marketing performance metrics for ray tracing into context. Claims for GigaRays per second may initially impress but will pale in reality due to low efficiency, linked to low utilisation of GPU processing resources or due to memory access limitations due to the non-coherent memory access patterns. How the hardware architecture helps to tackle these complex issues is explained in our white paper, so do check it out to learn more.

Once the ray tracing levels are understood, we can look forward to SoC designers enabling this game-changing technology to move from low efficiency, high-power consumption hardware into area and power-efficient devices, so mobile gamers will be able to enjoy the benefits of incredibly lifelike, photorealistic games for even greater levels of immersion.

Benny Har-Even

Benny Har-Even

With a background in technology journalism stretching back to the late 90s, Benny Har-Even has written for many of the top UK technology publications, across both consumer and B2B and has appeared as an expert on BBC World Business News and BBC Radio Five Live. He is now Content Manager at Imagination Technologies.

7 thoughts on “Introducing the Ray Tracing Levels System – and what it will mean for gaming”

  1. This question is off-topic but could you guys theoretically modify the drivers to allow tiles to render at different resolutions? The scenario where I’m seeing that being usefully is something like foveated rendering.

    I remember reading that the Oculus Go, which uses Adreno graphics, comes with a driver that allowed the GPU tiles at 1/2, 1/4th, 1/8th, or 1/16th resolution. If not, could something similar be done where 1×2, 2×2, 2×4, and 4×4 groups of tiles can be combined so that each frontend is still dealing with a 32×32 pixel?

    Reply
    • Hi Mark, thanks for the question. Yes, these are all concepts which are being considered and enabled in current/future designs – most of this falls under the umbrella of VRS – Variable Rate Shading.

      Reply
  2. Thank you for replying too sir.
    That double check the box then..

    What do you think if we add some concurrent computing & VRS
    into the next PowerVRs?
    Is it possible? And, is it beneficial?
    I’m sorry if i am being rude.

    Looking forward to B & beyond.

    Reply
  3. Hm, I think if we make the rasterization tiled before being implemented in Hybrid Rendering/Ray Tracing, it may increase the efficiency.

    Reply

Please leave a comment below

Comment policy: We love comments and appreciate the time that readers spend to share ideas and give feedback. However, all comments are manually moderated and those deemed to be spam or solely promotional will be deleted. We respect your privacy and will not publish your personal details.

Blog Contact

If you have any enquiries regarding any of our blog posts, please contact:

United Kingdom

benny.har-even@imgtec.com
Tel: +44 (0)1923 260 511

Search by Tag

Search by Author

Related blog articles

ray tracing feature image

Why benchmarks are the missing piece of the ray tracing puzzle

Today, as the industry works on fleshing out the accelerated ray tracing ecosystem on Windows-based PCs, consoles, and eventually, onto other markets in the future, (with bigger GPUs leading the way), there’s a missing piece of the puzzle that’s yet to show up: ray tracing performance analysis software.

Read More »

Connect

Sign up to receive the latest news and product updates from Imagination straight to your inbox.