shader model 4.0

[kalniel]

But when the testing is done - SM2 mode is forced - instead of letting each card run the shader model that it was designed for

Test everything with its 'natural' shader model + give marks for having features

I think I agree with this. While SM1 vs SM3 might show a faster result for SM1, the looks should be convincing enough for SM3's arguement. I think it's a result of some games going down the SM1 or SM3 route, with no SM2 in there for the ATI cards (this generation). Maybe SM2 is slower for the same things as SM3 as some developers have mentioned, but its inclusion would let the cards be tested on merit. And if ATI's implementation of SM2 is faster than the same effects using SM3 then let it be seen.

[Stoo]

tbh, I think it's more likely to be a sort of SM3.5, ie a few add-on bits from the xb360 development?

[Applecrusher]

I think I agree with this. While SM1 vs SM3 might show a faster result for SM1, the looks should be convincing enough for SM3's arguement. I think it's a result of some games going down the SM1 or SM3 route, with no SM2 in there for the ATI cards (this generation). Maybe SM2 is slower for the same things as SM3 as some developers have mentioned, but its inclusion would let the cards be tested on merit. And if ATI's implementation of SM2 is faster than the same effects using SM3 then let it be seen.

Practically every games uses SM2.. HL2, Doom3 (??).. everything...

[Stoo]

Yup, which was ATi's reasoning for not putting SM3 support in the x800 series, it saves silicon space by not implementing a function that (still) virtually nothing supports..

[zulander]

By buying a card that supports SM3.0 are you providing any level of future proofing> i.e. are more games going to use it in the future or is it just something that’s going to be ignored and surpassed and thus become irrelevant? Should same generation cards that support SM3 over a cards that do not be given any more brownie points or is not worth considering?

[Rys]

ATI are just about to release their own Shader Model 3.0 parts, so you'll be buying it by default if you buy a graphics card in the next year. If you're buying one now, you'd be mad not to at least consider an NVIDIA offering with that capability.

Shader Model 4.0 exists and has done since the inception of the latest D3D/WGF specification. A few minutes on Google would have told you a bit about the unification of vertex and fragment shader capabilities.

The graphics card market over the last few years should have taught you that a critical mass of parts is needed installed in PCs before developers shift efforts to supporting their feature set. That's there, or very close. Expect the SM3.0-supporting games to carry on increasing (and there's plenty already with Google again showing you list upon list of them).

As for this, which you've pushed a few times recently:

The concept of SM3 is that it allows 'really cool stuff' like texture fetches in the vertex engine and branching in the pixel shader...

...wouldn't it be funny if developers were steered away from using those features because existing implementations were just to slow

I'd stop throwing it around, for fear of it embarassing you in the future The Chronicles of Riddick doesn't support SM3.0 either, being an OpenGL game. Try Shader Model 2.0-class with soft shadowing via fragment programs instead...

And what's XDA? You mean XNA, which is just a toolkit/middleware? WGF and DX10 is next in the world of 3D APIs on Microsoft platforms.

ATI are working with MS on this one, but you can be sure nvidia will be playing catch-up real quick if it proves a good move.

Just because the API presents a unified caps and instruction set for both vertex and fragment units, doesn't mean the hardware has to have unified shader silicon. You can abstract it any way you like.

Silly thread with silly things being said!