3:8 vertex and pixel shader in X800 and NV40 - coincidence?

[Ailuros]

Talking about geometry, can they improve the triangle setup rate ? They seems to be only scaling setup rate with clock it seems, that's why all the focus on pixel shading rather than geometry.

I'd speculate that "more" would be theoretically always possible, yet size and cost are always a consideration too. If only the geometry unit(-s) (and not all other essential parts of an accelerator) would benefit from higher frequencies, then I guess it would be a different story.

In a recent (relevant) debate with a friend he pointed me at Sweeney's recent interview here at B3D, that might give an explanation why developers seem to be focusing more on pixel shading (for the time being of course):

Our pixel shaders in the Unreal Engine 3 demos are typically 50-200 instructions in length, and are composed from a wide range of artist-controlled components and procedural algorithms.

Our vertex shaders are quite simple nowadays, and just perform skeletal blending and linear interpolant setup on behalf of the pixel shaders. All of the heavy lifting is now on the pixel shader side -- all lighting is per-pixel, all shadowing is per-pixel, and all material effects are per-pixel.

Once you have the hardware power to do everything per-pixel, it becomes undesirable to implement rendering or lighting effects at the vertex level; such effects are tessellation-dependent and difficult to integrate seamlessly with pixel effects.

http://www.beyond3d.com/interviews/sweeneyue3/

Tesselation-dependent though could mean that WGF (think different topology/tesselation related philosophy), might change the focus to be more balanced between PS and VS.

 

[overclocked]

question to all here, is the VertexShader being overseen or not as used as it could be??

 

[Ailuros]

question to all here, is the VertexShader being overseen or not as used as it could be??

I'd rather dare to speculate that Vertex Shaders aren't flexible enough yet, to gain as much in importance as one would expect them to. It doesn't look like it's being overseen, rather the way Pixel shading is being used up to now and the so far capabilities of VS units, seem not to help much in getting a better balance between PS and VS.

I'm not even sure if I'm on the right track here, but when Sweeney mentions tesselation dependency, then for now it could mean that adding rendering/lighting effects to VS would add an unwanted added load on the host CPU.

 

[V3]

I'd rather dare to speculate that Vertex Shaders aren't flexible enough yet, to gain as much in importance as one would expect them to.

I guess so. Whatever happend to the elusive primitive processor, that gets thrown around here before the launched of NV40 or R420 ? Will we get to see it before CPUs becomes significantly faster making primitive processor redundant ?

 

[Ailuros]

I'd rather dare to speculate that Vertex Shaders aren't flexible enough yet, to gain as much in importance as one would expect them to.

I guess so. Whatever happend to the elusive primitive processor, that gets thrown around here before the launched of NV40 or R420 ? Will we get to see it before CPUs becomes significantly faster making primitive processor redundant ?

Look in the DX-Next preview here at B3D under the Topology/Tesselation sections of the article or check any newer presentations regarding WGF that refer to Tesselation units and/or Geometry Shaders.

A fully programmable primitive processor could of course take on both departments mentioned above. Trouble being that you need the necessary API support at least for WGF, since in OGL I guess proprietary extensions would solve that one.

PPPs on NV4x/R4xx? That was an endless myth regarding NVIDIA's plans and it had been also rumoured prior to release to be on the NV3x. Considering the API milestones above, I have severe doubts that such units will be considered prior to the WGF release or damn close to it.