Microsoft has announced DirectX Raytracing 1.2 at the GDC conference with claims of "game changing" performance benefits. But the newly revised API looks an awful lot like it has incorporated some key Nvidia technologies rather than introducing new features and so may not be the ray-tracing performance panacea we've all been waiting for.
Microsoft is claiming performance speed ups of 2x and more for DirectX Raytracing 1.2. Specifically, Microsoft says the introduction of support for opacity micromaps (OMM), "significantly optimize alpha-tested geometry, delivering up to 2.3x performance improvement in path-traced games. By efficiently managing opacity data, OMM reduces shader invocations and greatly enhances rendering efficiency without compromising visual quality."
The other key performance-enhancing feature is shader execution reordering (SER). Microsoft says it, "offers a major leap forward in rendering performance — up to 2x faster in some scenarios — by intelligently grouping shader execution to enhance GPU efficiency, reduce divergence, and boost frame rates, making ray-traced titles smoother and more immersive than ever. This feature paves the way for more path-traced games in the future."
- AMD, Intel, Microsoft, and Nvidia are all excited about cooperative vectors and what they mean for the future of 3D graphics, but it's going to be a good while before we really see their impact
- AI will be crammed in more of the graphics pipeline as Nvidia and Microsoft are bringing AI shading to a DirectX preview next month
Our understanding is that Nvidia RTX GPUs, at least back to the RTX 30-series, already support both of those features in hardware. SER support appears to have been added with the RTX 30-series, while OMM was there from the very beginning with RTX 20-series GPUs.
In other words, the claimed performance benefits Microsoft is talking about are already present in most Nvidia RTX GPUs. That is likely why Microsoft says, "Nvidia has committed driver support across GeForce RTX GPUs," but that it is, "actively working with other hardware vendors, including AMD, Intel, and Qualcomm, to ensure widespread adoption." In other words, RTX GPUs are already good to go.
Notably, a slide posted by Microsoft includes a quote from Intel saying that the company is , "looking forward to supporting SER when it is available in a future Agility SKD release, with OMM support coming in future hardware." AMD is name-checked by Microsoft, but there's no direct indication of its plans for OMM or SER.
What's more, it's important to understand the actual implications of these kinds of performance claims. The increases in performance of opacity micromaps or shader execution reordering (SER) likely would not boost overall path-tracing performance by the claimed amount, let alone actual frame rates.
Instead, those performance gains will accelerate the speed of that particular part of the rendering pipeline by the claimed amount. The rest of the time required to render a frame would be unaffected, including other elements of the path tracing pipeline.
All told what appears to be happening is more Microsoft adopting what Nvidia has already successfully implemented in its hardware. Now, that actually makes some sense. After all, it has been Nvidia pushing ray-tracing from the beginning and its RTX GPUs still have a clear performance advantage in ray-tracing and particularly in path-tracing.
The only slight snag is that a DirectX API built around Nvidia hardware, or at least following its lead, inevitably favours Nvidia GPUs when it comes to performance. And that's not necessarily great for market competition. Put another way, if DirectX tends to mirror Nvidia hardware specs, the likes of AMD or indeed Intel will always be one step behind.
Still, if ray-tracing is going to be a thing—and at this point it looks like a done deal—baking performance-enhancing hardware requirements into DirectX has got to be a good idea. AMD has certainly upped its game with the Radeon RX 9000-series of GPUs. But it would be no bad thing if Microsoft required further hardware features for AMD's next-gen GPUs.
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
A Jensen Huang-signed version of this golden Asus RTX 5090 will be auctioned off to support relief efforts for the California wildfires
Nvidia CEO sets sights on making 'several hundred billion' dollars worth of electronics in the USA over the next four years, increasing the chance of your next GPU being made in America
