Full NV35 official feature list inside

[Arun]

I've recieved, of course not in an official manner and I'm personally not under any NDA, the full NV35 feature list.
It's mostly PR talk, since it's pretty much the PR you'll get at launch, but it's still quite interesting.

I've used bold for everything I find noteworthy.

 CineFX ( TM ) Shading Architecture
 Support for the Microsoft DirectX® 9.0 Pixel Shader 2.0
 Support for the DirectX 9.0 Vertex Shader 2.0
 Long pixel programs up to 1,024 instructions
 Long vertex programs up to 256 static instructions with up to 65,536 instructions executed
 Dynamic, conditional execution and flow control
 256-bit advanced memory interface combines a wider memory data path with next generation controller technology for superior performance and throughput.
 Full 128-bit, studio-quality floating point precision through the entire rendering pipeline with native hardware support for 32 bpp, 64 bpp and 128 bpp rendering modes
 Accelerated pixel shaders allow for up to 12 pixel shader operations/clock
 Up to 16 textures per rendering pass
 Support for sRGB texture format for gamma textures
 IntellisampleTM performance technology, a Hi-Res compression technology (HCT), increases performance at higher resolutions through advances in compression and anti-aliasing technology.
 Shadow Volume Accelerator – accelerates shadow volumes by maintaining them more accurately and discarding useless information.
 Full nViewTM multi-display technology capability
 Integrated NTSC/PAL TV encoder supporting resolutions up to 1024x768 without the need for panning with built-in Macrovision copy protection
 Microsoft® Video Mixing Renderer (VMR) creates support for multiple video windows with full video quality and features in each window
 DVD and HDTV-ready MPEG-2 decoding up to 1920 × 1080i resolutions
 Dual, integrated 400 MHz RAMDACs for display resolutions up to and including 2048 × 1536 @ 85 Hz
 Dual DVO ports for interfacing to external TMDS transmitters and external TV encoders
 Dual internal TMDS encoders (one single and one Dual link) able to drive next-generation flat panel displays with resolutions greater than 1600 × 1200
 Digital Vibrance Control 3.0
 0.13 micron process technology for higher levels of integration and higher operating clock speeds
 40 mm × 40 mm, BGA 1309 flip-chip package
 Comprehensive Microsoft DirectX® 9.0 (and lower) and OpenGL 1.4 (and earlier) support
 DirectX and S3TC texture compression

Please also note nVidia is not marketing any DDR-II in this feature list. This seems to imply the NV35 *does* use DDR-I.

The 12 PS ops/clock thing is rather vague, because you could actually describe the NV30 that way too. It's still interesting to see nVidia won't market it as a "8 pipelines" architecture, though.

Sounds like the "accelerated shadow volumes" stuff is what Carmack was rumored to be asking about.

Also, I'd like to thank my source for giving me all of this very tasty info! And no nVidia, I won' tell you who it is. Don't be too angry though - not like I see a good reason to be angry since it doesn't really harm anyone's business


Uttar

EDIT: Thought the DVO thing was a typo, while it really wasn't. Fixed.

 

[Luminescent]

Thankyou very much Uttar.

What I don't see though are many new features. I wonder what kind of ops they're talking about by claiming 12 ops/clock. NV30 can do 8 FX in conjunction with 4 fp.

 

[Ozymandis]

If this is true, I only have two questions left-

When?

and

How will it compare to the 9800pro?

 

[LeStoffer]

Outstanding, Uttar!

The 12 PS ops/clock thing is rather vague, because you could actually describe the NV30 that way too. It's still interesting to see nVidia won't market it as a "8 pipelines" architecture, though.

Yes, it could be every thing. Maybe we are looking at 6 pipelines with a FP and FX path each. And since they are "Accelerated" they could made the wise decision to make separate FP texture and FP ALU units.

Sounds like the "accelerated shadow volumes" stuff is what Carmack was rumored to be asking about.

.. or it could be a marketing thing? nVidia already have hardware acceleration in the GeForce line while ATI don't AFAIR. Maybe the implementation is less coarse and thus more useful in the NV35 however.

Edit: I was thinking about support for shadow buffer on the GF3/GF4.

 

[Dave Baumann]

Sounds like the "accelerated shadow volumes" stuff is what Carmack was rumored to be asking about.

NVIDIA's HSR suffered from the same type of issues with stencils that R300 did. R350 was tweaked to fix that and it seems that NV35 has done the same.

 

[Luminescent]

How about fp shader performance?

 

[SvP]

Support for the Microsoft DirectX® 9.0 Pixel Shader 2.0
Support for the DirectX 9.0 Vertex Shader 2.0

Where did all those "+" go Should be at least 3 of them

 

[kars]

A couple of questions:

What is:

 Comprehensive Microsoft DirectX® 9.0 (and lower) and OpenGL 1.4 (and earlier) support

Comprehensive? Shouldn't it read "Full support"?....

Also:

Integrated NTSC/PAL TV encoder supporting resolutions up to 1024x768 without the need for panning with built-in Macrovision copy protection

Thanks for the inflexible tv-out which won't let us watch dvds on a television set...

 

[Evildeus]

Interesting. Thx for the information Uttar

 

[Arun]

If this is true, I only have two questions left-

When?

E3, May 14th ( might be the 13th or the 15th )
If you read www.notforidiots.com/GPURW.php , you'd have known that for some time ( I love plugging my website! )

Well, yeah, no new features it seem. I don't think the NV30 supported sRGB textures though, though the R3xx did.
Notice I did bold the "dual internal TMDS" stuff, because the NV30 could only do up to 1600x1200 with flat panel displays, and not beyond.

I'm very intrigued about the 12 PS ops/clock thing too. Lots of possibilities remain.

Note that I really don't know about final performance, since I don't know about final clock speeds... I really don't think the R350 would be a very serious competitor to it, though - also the rumored R390 could be more serious.


Uttar

 

[Luminescent]

The R350 would be a serious competitor to NV35 if it remains capable of only 4 128-bit fp fragment shader ops ops per clock (even if it can do a texture op in parallel), within official DX9, at least. Maybe on int apps the NV35 would dominate, but unless the clockspeed is significantly higher and the anti-alaising performance has drastically improved, I don't see NV35 significantly outperforming R350.

 

[Grall]

12 instructions/clock, that's for all 8 "pipes" or what?

And what of antialiasing? If it's the same shit as in the original FX, there's no reason to bother with the follow-up no matter what other bells and whistles it's got.

Also, standard DDR sure is a disappointing step back. If indeed it features DDR memory.


*G*

 

[Luminescent]

Thats also what I say about the AA, at least for current games.

 

[Arun]

Yeah, you're right, the NV35 would probably only beat the R350 when it can use FX12. But it can use it in both Doom 3 and Stalker already - so it's sure to dominate it in quite a few major games at least.

It indeed sounds like there's nothing new for Antialiasing, at least when it comes to this feature list. This is very bad news I guess. Even if it can beat the R350 with 4x AA thanks to its amazing raw memory bandwidth, it'll look a lot worse too...
Not even a single mention of Gamma Corrected AA - although that's the type of thing the customer wouldn't understand anyway since the GFFX already supports "Dynamic Gamma Correction" - but still, they'd probably have noted it...

It's obviously 12 instructions/clock for the whole PS. The question now is no longer what's the theorical maximum, the question now is how frequently it can achieve it though...


Uttar

 

[Ozymandis]

E3, May 14th ( might be the 13th or the 15th )
If you read www.notforidiots.com/GPURW.php , you'd have known that for some time ( I love plugging my website! )

Well, yeah, no new features it seem. I don't think the NV30 supported sRGB textures though, though the R3xx did.
Notice I did bold the "dual internal TMDS" stuff, because the NV30 could only do up to 1600x1200 with flat panel displays, and not beyond.

I'm very intrigued about the 12 PS ops/clock thing too. Lots of possibilities remain.

Note that I really don't know about final performance, since I don't know about final clock speeds... I really don't think the R350 would be a very serious competitor to it, though - also the rumored R390 could be more serious.


Uttar

Mid-May? That seems pretty soon to me. nVidia is really screwing over people who buy the FX IMO

 

[Dave Baumann]

Even if it can beat the R350 with 4x AA thanks to its amazing raw memory bandwidth, it'll look a lot worse too...

Problem there is that bandwidth isn’t within their control. Anyone can adopt faster RAM.

Not even a single mention of Gamma Corrected AA

a.) That would be a significant change to the way the AA operates in all likelihood, so it wouldn’t surprise me if it wasn’t on NV35.

b.) I notice ATI slapped a patent on it. Don’t know the scope of that patent though.

 

[Arun]

Problem there is that bandwidth isn’t within their control. Anyone can adopt faster RAM.

Good point... I think nVidia is planning a GDDR-II model ( the memory controller supports DDR-I & GDDR-II, according to several rumors ) to ripost to that.

Not even a single mention of Gamma Corrected AA

a.) That would be a significant change to the way the AA operates in all likelihood, so it wouldn’t surprise me if it wasn’t on NV35.

b.) I notice ATI slapped a patent on it. Don’t know the scope of that patent though.

The ATI patent: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=1&f=G&l=50&co1=AND&d=ptxt&s1='Gamma+Correction'.TTL.&s2=ATI.ASNM.&OS=TTL/

It's that one, right? It seems *very* general, not specific about Antialiasing. nVidia still got Dynamic Gamma Correction, so I guess they aren't affected by that patent - or maybe they're trying to think they aren't, hehe.

Yes, Gamma Corrected AA would likely be a too major change... Too bad, it's a so nice feature!


Uttar

 

[stevem]

Notice I did bold the "dual internal TMDS" stuff, because the NV30 could only do up to 1600x1200 with flat panel displays, and not beyond.

Not necessarily true. It's possible to drive a larger panel via 2 (or more) DVI outputs from the same card if supported. To the graphics card/OS it appears to be multiple monitors. You then configure their arrangement via the display control panel. It's good to see an advance on this kludgy solution.

 

[Dave Baumann]

Good point... I think nVidia is planning a GDDR-II model ( the memory controller supports DDR-I & GDDR-II, according to several rumors ) to ripost to that.

I’m unaware of the specific behind the memory controller, but I think DDR is actually most logical. Think about this – NV30 is currently the only DDR-II design they have, however that is configured with quad 32-bit banks for 128-bit width. IMO, its unlikely they would choose to use 8 32-bit banks for a 256-bit bus, because of both complexity and optimisations. So if they choose not to go with 8 32-bit busses they have the choice of designing a new DDR-II controller or lift a desing they already have. Given this is a refresh product the latter of those two options is probably most favourable. Take on board the fact that NV31 uses dual 64-Bit DDR busses IMO its quite likely they will further parallelise the NV31 bus to utilise a quad 64-bit crossbar (like R300/R350), and which is why I also feel it quite likely they will be using DDR.

It's that one, right? It seems *very* general, not specific about Antialiasing. nVidia still got Dynamic Gamma Correction, so I guess they aren't affected by that patent - or maybe they're trying to think they aren't, hehe.

No clue, Just had “Patented” in a presentation.

 

[Arun]

There, found the quote I was searching for...

Posted by MuFu, 28th of March 2003:

Ok - got some info back very quickly! Not much though...

The NV35 memory interface supports both DDR-I and DDR-II. DDR-I support is complete and works well, DDR-II support is questionable - I suspect they have run into problems with the 256MB boards and we may only see 256MB DDR-II NV3x's when Samsung's 2nd gen GDDR2 is widely available (meaning they can use 8x256MBit modules and do away with the stubs - that along with routing & power considerations should allow greater stability at higher clockspeeds).

It comes from the following thread:
http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=9094


Uttar