Why is Apple slow to implement Hardware based Ray-Tracing?

[exoticSpice]

AMD, Nvidia and Intel have it on their GPUs and now ARM itself is coming out with a HW RT GPU next year.
Hopefully A16 has it otherwise Apple is going to lag behind.

HW RT is really useful in Blender and games.

 

[BootLoxes]

Wait until M3. M2 is basically just a slight refresh of M1.

I am not saying it will come with M3, but this will be the next big change in architecture so there is a chance.

 

[evertjr]

Maybe they don't see demand for real time ray tracing currently (and I don't think RTX will be a thing this generation... it has been poorly implemented and in most cases its not worth the performance impact). If gaming interest starts to increase I can see they pushing in that direction. They are probably working in making it efficient enough to bring it to the iPhone and scale up to the Mac.

 

[stvkmco]

I would also bet on M3. On the API side Apple is deeply invested in raytracing and 3D in general.

There’s also the chance we’ll see RT silicon in the M2 pro chips.

 

[jav6454]

AMD, Nvidia and Intel have it on their GPUs and now ARM itself is coming out with a HW RT GPU next year.
Hopefully A16 has it otherwise Apple is going to lag behind.

HW RT is really useful in Blender and games.

It's a technology that is still being developed. Demand is low except for a few specific examples. Furthermore, RT tech is mostly used for games, which Macs are known for.

 

[leman]

Because it’s a non-trivial endeavor if you want to do it right. Nvidia is the only vendor currently that offers a competent RT implementation. Others simply offer a bandaid in form of hardware-accelerated intersection computation. But RT is much more than that. The real problem with RT on GPUs is not intersections but execution path and data divergence which kills performance on SIMD hardware. Nvidia’s RT implementation is in actuality a sophisticated approach to reordering memory accesses so that hardware resources can be utilized efficiently when doing RT workloads.

BTW, I have to laught a bit every time one mentions that ARM has their own hardware RT. Given how bad ARM GPUs are their RT performance is likely going to be worse than the A14 RT that Apple offered two years ago.

 

[Xiao_Xi]

Maybe they don't see demand for real time ray tracing currently

Hardware-based ray tracing would be a killer feature for Mac Studio because it reduces significantly rendering times. For instance, Nvidia Optix reduces rendering time in Blender by 40% over Nvidia CUDA.

AMD HIP vs. NVIDIA CUDA vs. NVIDIA OptiX On Blender 3.2 - Phoronix

www.phoronix.com

Which solution is best: ARM's Immortalis or Imagination's IMG CXT?

Immortalis-G715

Immortalis-G715 is Arm’s new flagship GPU for high performance and the first Arm GPU to introduce hardware-based Ray Tracing support on mobile.

developer.arm.com

IMG CXT GPU - Imagination

Ray tracing for mobile. Bringing high-performance, desktop-quality ray traced visuals to the palm of your hand.

www.imaginationtech.com

 

[Basic75]

Hardware-based ray tracing would be a killer feature for Mac Studio because it reduces significantly rendering times.

How many people render on a Mac Studio?

 

[leman]

It's a technology that is still being developed. Demand is low except for a few specific examples. Furthermore, RT tech is mostly used for games, which Macs are known for.

Given Apples significant investment in making Metal RT relevant for production renderers they could achieve a healthy performance improvement in this area with a competent hardware RT solution.

 

[leman]

Which solution is best: ARM's Immortalis or Imagination's IMG CXT?

Immortalis-G715

Immortalis-G715 is Arm’s new flagship GPU for high performance and the first Arm GPU to introduce hardware-based Ray Tracing support on mobile.

developer.arm.com

IMG CXT GPU - Imagination

Ray tracing for mobile. Bringing high-performance, desktop-quality ray traced visuals to the palm of your hand.

www.imaginationtech.com

Since there are zero GPUs currently in commercial products there is not much to say. From the presentation it seems that Immortalis uses low-effort hardware units to accelerate intersection computation. As I wrote above, that’s not enough to provide a comprehensive RT solution. They claim 300% faster RT compared to software implementation which sounds about right if you replace 6+ instructions running on a dead slow GPU by a single one. Of course, the performance is going to be laughable in any case. Again, you can already do real-time RT on iPhones for limited areas like soft shadows or slightly improved lighting with decent performance. ARM won’t do any better here. The hardware RT is just a marketing move - “look, we also have this latest flashy thing!”, it doesn’t cost them anything to implement and doesn’t do much.

Imagination has a flashy white paper with bold claims about how advanced their RT solution is, but it’s been almost three years and their tech is nowhere to be seen. It’s hard to take them seriously under these circumstances. Who knows, maybe some of this tech will make it into Apple GPUs.

 

[MrGunny94]

Metal RT seems the step in the right direction. plus with new macOS Ventura support of upscaling and etc it seems like the next 2 Architectures from Apple will have heavy focus on RT for sure.

 

[Xiao_Xi]

Imagination has a flashy white paper with bold claims about how advanced their RT solution is, but it’s been almost three years and their tech is nowhere to be seen.

Which GPU are you talking about? Imagination introduced the IMG CXT last November.

Imagination launches the most advanced ray tracing GPU - Imagination

IMG CXT brings desktop-quality visuals to mobile.

www.imaginationtech.com

Although it looks like a paper launch, Innosilicon introduced two GPUs based on IMG BXT last November.

Chinese GPU Maker, Innosilicon, Unveils Fantasy One Chip For Discrete Graphics Cards - Up To 32 GB GDDR6X Memory, Dual-GPU Solutions & 5 TFLOPs

Chinese GPU maker, Innosilicon, has unveiled their first discrete chip, the Fantasy One, which will be featured in a range of graphics cards.

wccftech.com

 

[leman]

Which GPU are you talking about? Imagination introduced the IMG CXT last November.

Imagination launches the most advanced ray tracing GPU - Imagination

IMG CXT brings desktop-quality visuals to mobile.

www.imaginationtech.com

Although it looks like a paper launch, Innosilicon introduced two GPUs based on IMG BXT last November.

Chinese GPU Maker, Innosilicon, Unveils Fantasy One Chip For Discrete Graphics Cards - Up To 32 GB GDDR6X Memory, Dual-GPU Solutions & 5 TFLOPs

Chinese GPU maker, Innosilicon, has unveiled their first discrete chip, the Fantasy One, which will be featured in a range of graphics cards.

wccftech.com

Doesn't matter much in practical terms if no GPU with this technology is shipping. IMG can claim to have the best RT implementation ever, and maybe they do, but why should I care if I can't buy a product based on this stuff?

 

[jav6454]

Given Apples significant investment in making Metal RT relevant for production renderers they could achieve a healthy performance improvement in this area with a competent hardware RT solution.

True, but again, for what specific purpose other than bragging rights? Not all games even support RT and again the market for hardware RT is niche.

 

[stevemiller]

apple has been actively contributing gpu rendering support to blender. they also showcase how their gpu's can hold large scenes in video memory for rendering as a part of their m1 pro/max/ultra marketing. saying its a niche they aren't interested in is kinda weird. bringing up game support feels like a fundamental misunderstanding of who the target market of a pro/max/ultra system even is. there are lots of smaller studios and freelancers who like using macs who would love a boost to their 3d productivity with features like this.

 

[leman]

True, but again, for what specific purpose other than bragging rights? Not all games even support RT and again the market for hardware RT is niche.

Apple is heavily investing into production rendering software, with contribution to Blender and other projects as well as working with various software developers. If they manage to shorten the gap they have with Nvidia, it would generate tremendous positive PR for Apple Silicon. I think the main point is that while the market is niche, it’s very impactful and influential. Renderer software is a standard to go benchmark tool for many hardware reviewers for example. It’s a bit like the Mac Pro - a niche product, but in the end a product whose success drives sales and recognition for the brand.

 

[Xiao_Xi]

How many people render on a Mac Studio?

Good question. We should also ask:
- What's stopping 3D artists from using Mac Studio?
- How long will it take Apple to resolve those issues?

 

[jujoje]

Good question. We should also ask:
- What's stopping 3D artists from using Mac Studio?
- How long will it take Apple to resolve those issues?

Depends very much on the workflow, apps and industry.

For my work (VFX) I can absolutely use a Mac Studio - the Max is faster than my iMac Pro on every metric and the Ultra chip twice as fast (Houdini / Karma) and rapidly improving as they iron out the bugs. There’s nothing holding me back in my day to day - it even runs decently in my MBA for prototyping stuff.

In terms of what they need to fix, I think for the market apple is targeting (freelance and boutique studios) particularly for tv and advertising faster raytracing would be the main thing, alongside getting the holdout software ported.

Give it a year and I think the things should be pretty good workflow wise. Whether artists actually bite is another matter (as Lemen said, raytracing has a halo effect - Apple get something impressive working would be a game changer in terms of perception).

Getting apple silicon good for rendering and 3D is something apple has invested a significant amount of time on with the various vendors and key to their AR strategy so feel fairly confident in their commitment these days.

 

[jav6454]

Apple is heavily investing into production rendering software, with contribution to Blender and other projects as well as working with various software developers. If they manage to shorten the gap they have with Nvidia, it would generate tremendous positive PR for Apple Silicon. I think the main point is that while the market is niche, it’s very impactful and influential. Renderer software is a standard to go benchmark tool for many hardware reviewers for example. It’s a bit like the Mac Pro - a niche product, but in the end a product whose success drives sales and recognition for the brand.

It would. However, ask yourself, in your daily use and professional use, how many times are finding yourself wishing you had RT hardware acceleration?

Don't get me wrong, it's a nice feature to have, but in the world of priorities, RT isn't one right now.

 

[singhs.apps]

It would. However, ask yourself, in your daily use and professional use, how many times are finding yourself wishing you had RT hardware acceleration?

Don't get me wrong, it's a nice feature to have, but in the world of priorities, RT isn't one right now.

Everyday ?

 

[jujoje]

It would. However, ask yourself, in your daily use and professional use, how many times are finding yourself wishing you had RT hardware acceleration?

Don't get me wrong, it's a nice feature to have, but in the world of priorities, RT isn't one right now.

Everyday ?

Until I realise how limited shading is in GPU renderers, and how much it would cost to render production scenes. My main hope with apple silicon is that they get gpu rendering fast enough that you can render larger scenes on the gpu. The amount of memory available is tbh more interesting that pure raytracing speed. Very different workflows and requirements I guess

Feature wise it’s getting closer but even there you’re limited by the subset of features possible on a gpu (point clouds, procedurals and volume sampling in shader tend to be unsupported).

 

[leman]

It would. However, ask yourself, in your daily use and professional use, how many times are finding yourself wishing you had RT hardware acceleration?

Not at all, I don’t need that stuff. I mean, I do some GPU graphics programming as a hobby and RT is fun to play with, but that’s about it.

Don't get me wrong, it's a nice feature to have, but in the world of priorities, RT isn't one right now.

Ah, but don’t discount user psychology. RT, overhyped as it might be, is the latest fashion. Even ARM GPUs donRT now, even if it’s a joke. Apple not providing hardware RT makes them look like they are behind in technology, and that’s a problem for them.

Besides, what other GPU features you would consider a priority right now? Apple is already ahead of the curve working y everything else, especially when you consider the software stack. A competent hardware RT solution might even bring benefits beyond RT as it needs some sort of sophisticated work item reordering unit. I’m only speculating at this point but maybe the same hardware can be used to make compute more efficient.

 

[Xiao_Xi]

It would. However, ask yourself, in your daily use and professional use, how many times are finding yourself wishing you had RT hardware acceleration?

Don't get me wrong, it's a nice feature to have, but in the world of priorities, RT isn't one right now.

It's not about how many times you use it, it's about whether 3D artists could make more money with a Mac Studio with hardware-based ray tracing and whether Apple would sell more Mac Studios.

What else do you want Apple to develop?

 

[EntropyQ3]

True, but again, for what specific purpose other than bragging rights? Not all games even support RT and again the market for hardware RT is niche.

Actually very few games outside some poster children sponsored by Nvidia does.
Reality check - this is a list of the highest grossing games on iOS. A huge part of this list isn't even 3D-rendered. Of those that are, no-one in their right mind would suggest that what they need is for instance RT reflections because screen space reflections would be unacceptable. They are not even remotely close to a level of rendering fidelity where limitations of lighter weight lighting methods are relevant. Arguably the limitations of mobile platforms lie entirely elsewhere. Nevermind that the staggering majority of the billion or so iOS users just don't care how lighting is calculated on their phones and tablets. So is dedicating hardware resources to RT on Apple GPUs an intelligent way to use the available transistor budget? We have a good idea what the future might bring in terms of transistor budget within the next half decade, and the short answer is - not a lot.

So is it worth it for Apple to add dedicated RT hardware exclusively for the benefit of those few who use Macs for professional rendering (as opposed to benchmarketing using Blender)?

Is it a sufficiently selling buzzword that Apple requires a ticked marketing checkbox?

I'd rather see Apple spending their transistor budget on features that has a wide immediate usage or wide applicability. If that went into 3D graphics at all, then spending them on features that for instance lighten the load in terms of asset storage, or more efficient high quality upscaling seems a lot more useful.

 

[leman]

So is it worth it for Apple to add dedicated RT hardware exclusively for the benefit of those few who use Macs for professional rendering (as opposed to benchmarketing using Blender)?

Markets change, and it's not like a competent RT requires a massive transistor budget. If Apple were to deliver useable real-time RT acceleration it would make them the first company offering this feature on a reasonable power budget (unlike Nvidia where you need a 200W desktop GPU for RT in games to make any sense). That would truly bring raytracing to the masses (at least to the masses of Apple users) and put Apple in a unique position.

I'd rather see Apple spending their transistor budget on features that has a wide immediate usage or wide applicability. If that went into 3D graphics at all, then spending them on features that for instance lighten the load in terms of asset storage, or more efficient high quality upscaling seems a lot more useful.

Apple already has hardware compression for both data and bandwidth and it's not like the upscaling requires any dedicated hardware. As I wrote before, a good hardware RT implementation has to start with a comprehensive work/memory access reordering solution, which will have benefits way beyond RT alone. Getting good performance of the RT is not just about computing ray-triangle intersections quickly, but first and foremost solving the problem of control flow and data divergence. If this problem can be solved, the GPU suddenly becomes a much more competent programmable processor.