no 6800ultra Extreme, still the same 6800ultra

[radar1200gs]

Can you provide finalised specs for SM4.0? IIRC SM4.0 is a work in progress which makes it rather difficult for anyone to target at the moment, whereas SM3.0 has existed pretty much in its current form (with maybe a tiny tweak here and there) since DX9 was first released.

And you don't think NV, ATI, and Microsoft (along with several game developers) are working on this specification right now? Funny, that.

Dave was trying to imply that SM3.0 is just another waypoint on the DX9 roadmap. I'm saying it is (and was) the end of the currently defined roadmap (of course the roadmap might get extended, but I'd expect that then you would have a DX9.1 and SM4.0 would be like PS1.4).

I don't agree .


sm4.0 is obviously at the final stages. Ms is going to want something much more advance for the xbox 2 than an already 2 year old spec .

If anything this 9.0c dx is going to be like dx 8.1 . A small step that only one hardware producer is gonig to support .

Irrelevant. SM4.0 does not exist in DX9 yet. SM3.0 does and always has.

 

[KimB]

While I do expect SM4 for the X-Box, I expect SM3 for ATI's next architcture.

 

[ANova]

While I do expect SM4 for the X-Box, I expect SM3 for ATI's next architcture.

And I'm sure NV's next architecture will be SM3 once again.

 

[KimB]

Next architecture? No. Refresh? Yes.

 

[ANova]

Next architecture? No. Refresh? Yes.

The NV45 is a refresh, the NV50 is not. You can count on the NV45 launching at the same time as the R480

 

[KimB]

Right, and I expect the NV50 to offer quite a bit more than SM3.

 

[jvd]

Irrelevant. SM4.0 does not exist in DX9 yet. SM3.0 does and always has

RIght and when ati was supporting p.s 1.4 which existed in dx 8.1 . Nvidia didn't

p.s 1.4 support was very slow to take off and never really did.

So if we look at it this way.

sm 3 will not take off because ati doesn't support it but when dx 10 or dx next comes out the first set of shaders will take off .

History has taught us that .

 

[ChrisRay]

I'd have to say Pixel Shader 1.4 support has taken off some. I mean Nvidia uses it whenever they can get away with it as oposed to Shader 2.0

 

[jvd]

I'd have to say Pixel Shader 1.4 support has taken off some. I mean Nvidia uses it whenever they can get away with it as oposed to Shader 2.0

that will stop quickly once the 6800s actually hit the streets though. Nvidia will stop investing so much time to bring the nv3xs to usable lvls and spend that time trying to get the 6800s on parity with the x800s .

 

[ChrisRay]

I know Jvd, I was making a small joke too lighten up the conversation. I dont think Pixel Shader 1.4 is as dead as you guys believe tho,

There are several titles that use it atm. Its a pretty useful shader model. Specially if you dont need high precision shaders.

 

[ANova]

Right, and I expect the NV50 to offer quite a bit more than SM3.

That depends on when NV and ATI decide to launch their next gen cards. If it's before Longhorn then the R500 will probably have SM3 and the NV50 will still have SM3, as there will be little more that they can offer. If both decide to launch after Longhorn then both will probably support SM4.

 

[radar1200gs]

Irrelevant. SM4.0 does not exist in DX9 yet. SM3.0 does and always has

RIght and when ati was supporting p.s 1.4 which existed in dx 8.1 . Nvidia didn't

p.s 1.4 support was very slow to take off and never really did.

So if we look at it this way.

sm 3 will not take off because ati doesn't support it but when dx 10 or dx next comes out the first set of shaders will take off .

History has taught us that .

Actually I'd say that history shows that the market supports what nVidia supports, and doesn't support what nVidia doesn't support...

 

[jvd]

Irrelevant. SM4.0 does not exist in DX9 yet. SM3.0 does and always has

RIght and when ati was supporting p.s 1.4 which existed in dx 8.1 . Nvidia didn't

p.s 1.4 support was very slow to take off and never really did.

So if we look at it this way.

sm 3 will not take off because ati doesn't support it but when dx 10 or dx next comes out the first set of shaders will take off .

History has taught us that .

Actually I'd say that history shows that the market supports what nVidia supports, and doesn't support what nVidia doesn't support...

Well we could say that about 3dfx too. But look what happened to them. Are you saying nvidia is the next to fail in this market ?

 

[radar1200gs]

They haven't yet, and if they were going to fail NV3x is probably what would have caused it.

EDIT: Besides, 3dfx's entire problem was that the market abandoned support for them. Hardly the case where nVidia is concerned.

 

[Ailuros]

Right, and I expect the NV50 to offer quite a bit more than SM3.

Hopefully yes; the only other remaining question mark would be API support after that. Unless Microsoft changes it's plans for any possible intermediate API beyond dx9.0, what I'd have in mind past SM3.0 would be - until API support is being supplied - be something more like checkboard features or OGL/proprietary extension material.

For the record didn't you expect the same also for NV40 some time ago?

 

[Ailuros]

On feature support in general: it's never ever enough if one major IHV supports a specific feature. In the majority of cases it takes a wider range of market acceptance for a specific feature to start getting utilized in games.

There's always of course a chance that it might get accepted by more than one IHVs down the road, but then it gets usually rendered redundant in relative terms by more advanced techniques. One good example that comes to mind is EMBM. By the time NV supported it in NV2x, we already had pixel shaders on GPUs.

In my book IHVs should continue to innovate and take their risks, irrelevant of the result.

 

[ChrisRay]

On feature support in general: it's never ever enough if one major IHV supports a specific feature. In the majority of cases it takes a wider range of market acceptance for a specific feature to start getting utilized in games.

There's always of course a chance that it might get accepted by more than one IHVs down the road, but then it gets usually rendered redundant in relative terms by more advanced techniques. One good example that comes to mind is EMBM. By the time NV supported it in NV2x, we already had pixel shaders on GPUs.

In my book IHVs should continue to innovate and take their risks, irrelevant of the result.

Enviromental Bump mapping is useful though, it can be used with pixel shader to create some great effects, Actually many games I know of use EBM.

 

[volt]

I thought they [ATI and NVIDIA] were supposed to slow down a notch?

Anyway I think Microsoft might not be able to keep up with newer shader models

 

[jvd]

They haven't yet, and if they were going to fail NV3x is probably what would have caused it.

EDIT: Besides, 3dfx's entire problem was that the market abandoned support for them. Hardly the case where nVidia is concerned.

well 3dfx put out the voodoo 3 and that was equal to the nv3x series . The nv40 will be just like the voodoo 5 and the nv50 will just never come out .


Of course nvidia can just keep paying devs to hold back support of advance features to keep all those who bought video cards from them for the last 2 years happy. But they wont. They will drop nv3x tech and optimzations and then push nv40 optimizing

 

[KimB]

For the record didn't you expect the same also for NV40 some time ago?

I don't remember expecting much about the NV40, but I think it seemed obvious that it would, at the very least, have longer instruction counts than SM3 requires at a minimum (since they already supported more instructions....). I was really surprised at the level of feature support the NV40 had, though, particularly the FP blending and filtering.