XBOX 2 GRAPHICS DETAILS EMERGE .. not much actually

+ implies unified architecture to me, but I bet you mean something between SM3.0 and SM4.0. Not sure what that means exactly.

Tommy McClain
 
I guess Xbox 2 graphics should still go somewhat beyond SM 3.0, if not significantly beyond 3.0, as that is what we've seen on the web previously....

http://editorials.teamxbox.com/xbox/858/The-Xbox-2-Inside-and-Out-Part-I/p3

A previously leaked document claims the Xbox successor will support Shader Model 3.0 and beyond, which can be interpreted as some of the features we’ll see in the next version of DirectX, set to be shipped with Windows Longhorn.

http://www.xbox-scene.com/xbox1data/sep/EplZAyukEVDWcUicJE.php
The Xenon graphics architecture is a unique design that implements a superset of Direct3D version 9.0. It includes a number of important extensions, including additional compressed texture formats and a flexible tessellation engine. Xenon not only supports high-level shading language (HLSL) model 3.0 for vertex and pixel shaders but also includes advanced shader features well beyond model 3.0. For instance, shaders use 32-bit IEEE floating-point math throughout. Vertex shaders can fetch from textures, and pixel shaders can fetch from vertex streams. Xenon shaders also have the unique ability to directly access main memory, allowing techniques that have never before been possible.

I too expect something inbetween SM 3.0 and SM 4.0

certain features in Xbox 2 VPU probably won't show up on the PC side until R600, if ever.
 
I also thought that Xbox was limited to PS !.3. Is there something about NV2A none of us know or were you simply making a typo and meant to put Xbox 2? Cuz if Xbox 1 has SM 3.0 and more then I would sur elike to know where!
 
I´m thinking more in terms of features, than chip name.

This chip compared to NV2A (relativily/
it has more FX ALU's, an extra VS (IIRC) and NV30's ROPs.
) dont seems so good.

Or ,at least, I expected to they use their experience to make something even better.

Dont you agree?
 
DaveBaumann said:
XBox = SM3+

OK, Xbox2's "SM3.5" may possibly work with unified shader model in the underlying hardware. I guess I just made too big an assumption, that because the article was talking about coding in SM3.0, that the Xbox2 hardware would be limited to SM3.0 defined hardware.

Thanks guys for clearing that up. :)
 
DaveBaumann said:
XBox = SM3+

BTW... will the R500 or whatever a discrete chip or somewhere hidden in the architecture, like being integrated in a north/south bridge ?

Actually I'm a bit surprised nobody started any speculation on the security stuff in the Xbox2 aka "Xenon"....

From what I know/hear people like "Bungie" (Hacking the Xbox) will have a hard time.
 
:?:

1. Are these already final chips or pre-production? i guess the later.

2. Do we know the clockspeed?

Any developers here have a devkit with r500 inside and would like to comment ;)
 
Sonic said:
I also thought that Xbox was limited to PS !.3. Is there something about NV2A none of us know or were you simply making a typo and meant to put Xbox 2?

Errr, yes that was a typo!!
 
ultimate_end said:
OK, Xbox2's "SM3.5" may possibly work with unified shader model in the underlying hardware. I guess I just made too big an assumption, that because the article was talking about coding in SM3.0, that the Xbox2 hardware would be limited to SM3.0 defined hardware.

I've said this before, but the best I can follow the architecural line is:

Code:
R300-->R420-->R520
               ^
               |
               |
       R400-->R500-->R600

The "R300" architecural line will basically end up as the DX9 desktop line, and end with R520 as the PC move over to Longhorn. R400, I believe, was always designed as the unified shader model however, would have been under powered, but over-featured for the PC when it was due - this will become the basis for the XBox2 (R500) and have the rest of the DirectX Next features tied up and be released as the R6x0 family on the PC for Longhorn. I believe that R500 is being termed as "SM3.0+" becuase there are still some shader length limitations in there (although we are talking thousands of instructions) whereas DX Next is fully virtual - my expectation would be that R500 would be closer to DX Next than SM3.0 but the limitations may prevent it from going to full DX Next (although DX Next, sorry, WGF may still move on at this point).

Note that I've put a line between R500 and R520 because although 520 is a heavy extension on the R300 line it will almost certiany implement quite a lot of work that has gone into R500's design.
 
Nvidia already was no shader length limitations at some time and if r400 is, originaly, to be out at the same time than 6800 it is not strange that r400/r500 is built that in mind, right :?:
 
R400 was orignally schedued to come back in July 2003, IIRC, which would have been a winter release instead of R360 (9800 XT).
 
DaveBaumann said:
Note that I've put a line between R500 and R520 because although 520 is a heavy extension on the R300 line it will almost certiany implement quite a lot of work that has gone into R500's design.
Yes that did get me wondering, as many people expect that R520 is essentially nothing more than an R420 with full FP32 precision.
But that would seem like a waste of a product launch.

Anyway. As you said, the Inquirer doesn't know the whole story.
 
pc999 said:
Nvidia already was no shader length limitations at some time and if r400 is, originaly, to be out at the same time than 6800 it is not strange that r400/r500 is built that in mind, right :?:

Since when have Nvidia been without shader length limitations :?:
 
OICAspork said:
pc999 said:
Nvidia already was no shader length limitations at some time and if r400 is, originaly, to be out at the same time than 6800 it is not strange that r400/r500 is built that in mind, right :?:

Since when have Nvidia been without shader length limitations :?:

In 6800 pdf I think, I may be wrong but Dave B. would correct me.

In other point why those reports never mention the + general programing features that DeanoC talked in a old thread :?:
 
Back
Top