Unreal Engine 5, [UE5 Developer Availability 2022-04-05]

Primitive shader most likely takes advantage of the ALU in the CUs.
If the Primitive shader is only taking advantage of the instruction set available to the CU and nothing else, then this doesn't make sense to me because you might as well run your shader code from your Compute shader, and test for a performance path output to the FF rasteriser when performance is required from them. Compute shader or Primitive Primitive shader, they both have access to CU instructions. Unless your Primitive shsder has access to additional instructions from a set of ALUs which may also include CUs, then these additional instructions made available to your Primitive shader makes switching worthwhile.

Anothet thing that doesn't make sense to me is that if both Epic and Sony were in discussions about future rendering technologies, and Cerny had access to AMDs IP portfolios, why didn't they design an advanced hardware Programmable Primitive block that would act as a fast function 'Reyes Rasteriser' processing micropolygons - similar to how traditional FF rasterisers act today but designed for REYES? This should leave the compute heavy work to all your CUs to focus on lighting and shading for Lumen.

Tim Sweeney has been thinking about REYES and voxels since at least as early as 1999. Below are his predictions from then for circa 2006/7:

https://techreport.com/news/46/sweeney-on-the-future-of-3d-graphics/

"2006-7: CPU’s become so fast and powerful that 3D hardware will be only marginally benfical for rendering relative to the limits of the human visual system, therefore 3D chips will likely be deemed a waste of silicon (and more expensive bus plumbing), so the world will transition back to software-driven rendering. And, at this point, there will be a new renaissance in non-traditional architectures such as voxel rendering and REYES-style microfacets, enabled by the generality of CPU’s driving the rendering process. If this is a case, then the 3D hardware revolution sparked by 3dfx in 1997 will prove to only be a 10-year hiatus from the natural evolution of CPU-driven rendering."

I can excuse his CPU to be modern day Compute in GPUs today. His REYES became Nanite, and voxels became Lumen. I suppose RDNA2 will be like a crossroads, allowing traditional rendering to exist with emerging alternatives.
 
@j^aws I'm sure there are some advantages to primitive shaders over compute shaders when you're processing vertices, otherwise it wouldn't exist. You're probably right about that. But they're designed to replace a particular part of the render pipeline to fit within the context of Direct3D and Vulkan. So they take vertices as inputs and output vertices, as far as I can tell.
 
And there we have it. For all we know the demo could have hovered around 30-45 on PS5 if uncapped. PS5 is 2070 Super level mostly I reckon.
They are aiming 1440p 60fps so surely they already reach 60 fps in many parts. I would bet when Lumen kicks heavily in is when the frame rate suffers, so thats way they are even considering mixing it with RT hardware.
 
And there we have it. For all we know the demo could have hovered around 30-45 on PS5 if uncapped. PS5 is 2070 Super level mostly I reckon.
The dynamic res tells us that is not the case, it was 1440p for most of the Demo with a drs on targetting 4K. That means the Gpu, even at 30 fps, was being fully utilised. I think epic was so clear about the Real resolution of the Demo because they have an engine to sell and they want Potential dev teams to know how it runs.

At least that is how dynamic res works in every Single game on Unreal Engine we have ever tested. That is why you use, to Max the Gpu and always get the best Image quality possible.

Sweeneys quote regarding dramatically different frame times in the Demo does not really align with DRS usage in Unreal Engine. You can test it yourself if you want to in Gears 5 or in almost any game using Unreal Engine unlocker. When you target an overly greedy resolution like 4k and apply DRS, it will keep frame times much more similar to eachother. It being mostly 1440p as epic said before would mean that many of the Demos frames cost a similar amount even without drs. If DRS has a stable res then that means the content has a stable cost.
 
Last edited:
I think he is in full damage control mode now, how can the video achieve 40fps? videos don't go faster than the source, he himself stated that RTX 2070 Super runs the demo with pretty good performance couple with an NVMe drive.

He seems confused, the event did have the video playing on a laptop for demonstration purposes, so he assumed people are asking about that, when -in fact- people are asking him about what the engineer said!
 
Last edited:
And there we have it. For all we know the demo could have hovered around 30-45 on PS5 if uncapped. PS5 is 2070 Super level mostly I reckon.

maybe in this demo (without HW based raytracing) ps5 is on level of RTX 2070 (not super, RTX2080 mobile is 1-4% more faster than a desktop rtx 2070)
but RTX have more silicon, it has raytracing units, that is not used by this demo, it has tensor cores, etc. we need a better benchmark, a tech demos using raytracing, dlss 2.0 etc, maybe even an RTX 2060 can outperform ps5, it's too early to compare, we need better tools/testbed. And the best testbed will be the games, not side-pay techdemo made to shine in one single platform

we should wait to compare perfomances, in this moment
 
The dynamic res tells us that is not the case, it was 1440p for most of the Demo with a drs on targetting 4K. That means the Gpu, even at 30 fps, was being fully utilised. I think epic was so clear about the Real resolution of the Demo because they have an engine to sell and they want Potential dev teams to know how it runs.

At least that is how dynamic res works in every Single game on Unreal Engine we have ever tested. That is why you use, to Max the Gpu and always get the best Image quality possible.

Sweeneys quote regarding dramatically different frame times in the Demo does not really align with DRS usage in Unreal Engine. You can test it yourself if you want to in Gears 5 or in almost any game using Unreal Engine unlocker. When you target an overly greedy resolution like 4k and apply DRS, it will keep frame times much more similar to eachother. It being mostly 1440p as epic said before would mean that many of the Demos frames cost a similar amount even without drs. If DRS has a stable res then that means the content has a stable cost.
OK sure thing. But nowhere in that Chinese podcast did they mention about the laptop running a fixed 1440p res, for all we know it could also be running with DRS on, maybe not even at 1440p at all. The nature of that laptop performance was very vaguely presented so I don't think we can rely too much on that for now. More so that PS5 demo was running on early devkit and they were aiming for 60fps, who knows if the final optimized build could put it on the same level as a 2070s. But I admit I jumped into conclusion too fast myself before, we should wait for more data.
 
I love how everyone on this forum understands UE5 better than the founder and CEO of the company that makes it.
Well, I wouldn't expect a typical CEO to know anything about their products; they tend to have other concerns and hire people to work on engineering products. A CEO still hands on with the technical implementation of their stuff is fairly exceptional. So TBH, people here being better experts on something over the CEO of the multinational that makes it isn't that unrealistic. ;)
 
The dynamic res tells us that is not the case, it was 1440p for most of the Demo with a drs on targetting 4K. That means the Gpu, even at 30 fps, was being fully utilised. I think epic was so clear about the Real resolution of the Demo because they have an engine to sell and they want Potential dev teams to know how it runs.

At least that is how dynamic res works in every Single game on Unreal Engine we have ever tested. That is why you use, to Max the Gpu and always get the best Image quality possible.
Is there any reason to lock upper resolution and just use DRS for low resolutions to maintain framerate? Technically that's possible. Perhaps 1440p is the limit of part of the pipeline on PS5 so they can't render above that? Dunno, but as this engine isn't doing the same as other UE titles, we can't really use those as a basis for comparison. We need a better argument based on technical reasons rather than precedent (of which there is none ;)).
 
I believe you are misunderstanding how this works. The output of primitive shaders feed the rasterizer hardware. The UE5 engine software rasterizes and bypasses the raster hardware. The majority of the triangles are software rasterized. The minority of triangles, where the hardware rasterizes would be faster (probably related to cases where polygons are bigger than one pixel), the engine falls back to primitive shaders instead of the old vertex shader pipeline. For the majority of the scene the primitive/raster units are not used.
Are you sure primitive shaders can't handle virtual triangles? It seems like they're saying they'll fallback to software CUs, when the adequate hardware is not there and thus is slower. I've heard the functionality of the primitive shaders compared to that of the vector units of ps2 which should definitely be able to handle virtual triangles.
No, it won't, they clearly state the demo doesn't rely on the ray 5.5GB speed of the PS5 SSD, and can run on a regular PC SSD.
The laptop is also a high end laptop with nvme. Some translations suggest he said a decent ssd could handle it, would like to see what patsu clarifies regards that.

Even with this high end laptop it seems they needed to optimize layout on ssd for fixed flight path which wont work in open ended flight path.
Primitive shader most likely takes advantage of the ALU in the CUs. The fixed-function raster hardware in the shader array does not, at least not to my understanding. It's a special purpose unit that assembles triangles from vertices, culls triangles and outputs pixels from coverage tests. Each unit can rasterize 1 polygon per clock, outputting up to 16 pixels. Both PS5 and Xbox Series X have 4 shader arrays, so 4 raster units each. The Series X has more CUs per shader array, but the performance of the rasterizer is still 16 pixels from 1 polygon per clock.

I don't think the rasterizer cares where it gets the vertices from (vertex shader, primitive shader, mesh shader, compute shader). The difference in UE5 is they don't use the rasterizer hardware at all for the vast majority of the pixels on screen. They replace its function in the compute shader to get better performance for small triangles. The compute shader accepts data, probably vertices and outputs pixels.
If the primitive shaders involve a rework of h/w might they not be faster than the compute shader route?

[mesh shaders]
NVidia developer resources imply it's actually a new geometry pipeline which supersedes the traditional vertex, tessellation (hull/domain), and geometry shader stages - that would require a lot of work on the supporting A-DmitryKo
If that were so, would the hardware be part of geometry handling hardware in the gpu? or would it be part of the CUs? When someone like sony says geometry engine, do they mean a specific piece of hardware or simply a modification to CUs or some software solution?


On the subject of laptop GPU ...

The engineer said Lumen uses up more resources than Nanite. Although their target for nextgen consoles is 60fps, he hasn't hit it yet.
He was going to guarantee that he will [hit 60fps]. But stopped short of completing his statement and then as a matter of fact, said that he already got laptops to run at 40fps near the beginning of the demo in the editor. [He didn't mention the resolution, effects, and general settings for his test]


.
I wonder if that version running at 40fps is the same as the lower detail version that runs at 60fps in some hardware(probably ps5). Or if he hinted that it was more detailed. Did he say our 60fps optimizations can currently get even 40fps on a laptop or something to that effect?
 
Is there any reason to lock upper resolution and just use DRS for low resolutions to maintain framerate? Technically that's possible. Perhaps 1440p is the limit of part of the pipeline on PS5 so they can't render above that? Dunno, but as this engine isn't doing the same as other UE titles, we can't really use those as a basis for comparison. We need a better argument based on technical reasons rather than precedent (of which there is none ;)).
What do you mean by your first question, could write it out with numbers?
I do not understand
 
Back
Top