"Some sort of ATI graphics card"

[LeStoffer]

To me the words "at the moment is the fastest platform to run the game on" to indicate that R300 is in a closer state of readiness, but the performance title is still up fr grabs!

I'll second that. It is quite a turn of events from the DX8 lanch days where nVidia was half a year faster (GF3 vs R8500). I just wonder why ATI could cut it faster? I think that it boils down to the complexity of design, e.g. nVidia had to scrap more of their knowledge from GF3/GF4 than ATI had to do from R200.

Ahhh, the pure pleasure of speculation!

 

[ben6]

The hardware that ran Doom3 is definitely ATI's next generation. And it WAS impressive. I'm sorry , I can't go much into my Nvidia meeting (nice booth with UT2k3, Project Gotham Racing Earthweb and a pixel/vertex shader version of Counterstrike for the XBox) , but it is interesting what Nvidia said about Matrox. Here's the quote:

It's a interesting piece of hardware with some interesting new features. The problem as I see it is, it's still a DX 8.1 part that's releasing very near the fall. In the fall Microsoft , as you know will launch DX9.

Actually a very fair statement IMO. If it seems like I cut out something you'd be correct

 

[Typedef Enum]

I'm just guessing here...But doesn't it really seem like NV30 is truly going to be a departure from any/all previous offerings?

I mean...If it really was nothing more than a NV25 with DX9 functionality, it wouldn't seem to have any serious impact on development.

Given the statements made over the last 2 months or so...The CEO making his comments, David Kirk's statements, Carmack's, and the fact that most people knew/expected that 3dfx technology would, in some way, be thrown into the mix...It does seem as if we should be prepared to expect...the unexpected, as far as the tech. that we've seen to date from nVidia.

I think it's definitely a good thing that R300 is at such an advanced state right now. It will certainly put nVidia "on notice," if it hadn't happened already.

Yes...I also believe ATI directly responded to my post

 

[Typedef Enum]

Ben, did you by any chance get any sort of feedback/comments from them with regards to ATI?

 

[Joe DeFuria]

I mean...If it really was nothing more than a NV25 with DX9 functionality, it wouldn't seem to have any serious impact on development.

Well, I think saying "NV25 with DX9 functionality" doesn't make much sense. Seems to me that the DX9 pixel pipeline is a pretty radical departure itself from the DX8.0 pixel pipeline.

On the other hand, DX9.0 (PS 2.0) seems MUCH closer to PS 1.4 than PS 1.1-1.3. In other words, if the R300 and NV30 turn out to be "functionally equivalent", they will be much closer to R200 architecture than NV25.

In short, ATI has a less radical departure going from R200 to R300, than nVidia does going from NV25 to NV30, IF NV30 is essentially equivalent to R300. So it doesn't really surprise me that the R-200 to R-300 development cycle is much shorter than the Nv20 to NV30 development cycle. It also won't surprise me, if ATI's drivers are much more stable OOTB.

 

[Typedef Enum]

I wasn't trying to imply that DX9 wasn't much of a departure from DX8...I was only trying to convey the possibility that nVidia was, indeed, making a radical departure from their previous offerings...We can say that the NV2x chip was a big step-up from, say, the GF256/GTS...but at the same time, there was a lot of inheritance as well.

I just have the feeling that NV30 will, indeed, be an enormous departure for nVidia...what that is, however, isn't very clear at all.

 

[Jim Macintosh]

Perhaps I am mistaken, but isn't ATi the hardware vendor working closely with MS to develop DX 9 (much as NVIDIA did for DX 8)? If this is correct, then the advanced state of R300's development is a bit less of a surprise. ATI will have a much better idea of exactly what DX 9 entails.

That being said, I also think that what some of you are saying is also correct: R300 may be less a departure from previous technology than NV30.

Is the R300 .13 technology? If so, this would suggest that availability of this process is a bit less problematic than perhaps some have suggested and is not a direct reason for NVIDIA being a bit slower to get their chip ready.

Anyway, I would be very careful inferring too much from any of this. Untill there are more substantial and credible indications of what NVIDIA actually is up to with the NV30, it is idle to speculate about the relative merits of the four or five new architectures.

 

[Joe DeFuria]

Yes, I know what you (typedef) are postulating.

What I'm trying to say is that ANY DX9 compliant part itself, is a radical departure from any DX8 or lower part. No "3dfx tech", or special unforseen tech, or any such thing is required, for anyone to call their new DX9 part "a radical departure", "our most significant contribution" to the industry.

I just think that many nVidia supporters are reading too much into nVidia CEO and marketing statements about their next big thing. Supporting DX9 itself IS a big thing.

It is an "interesting" spectacle to see the nVidia sites react to the whole DoomIII & R300 announcement.

 

[Joe DeFuria]

According to the leaked ATI road-maps (which are months, if not a year old at this point), ATI had the R300 slated for a 0.15 micron process, 8 pixel pipes and DX9 compliant. However, I tend to think such a part is not possible. I'm expecting any DX9 part (R300 and NV30) to be produced on 0.13.

There may still be problems with 0.13. We'll find out when we see the prices and general availability of these parts this fall.

 

[Dave Baumann]

Is the R300 .13 technology? If so, this would suggest that availability of this process is a bit less problematic than perhaps some have suggested and is not a direct reason for NVIDIA being a bit slower to get their chip ready.

I've heard it bother ways - .13um and .15um so I don't think we're really going to know until its released.

Given the proximity of .13um being a viable prospect I think performance wise if you don't bring out something on .13um you are likely to be suprassed by your competitors pretty soon when they start using it. Also you have to question if a .15um architecture can really offer all that much more performance and DX9 functionality.

 

[Geeforcer]

Yes, I know what you (typedef) are postulating.
It is an "interesting" spectacle to see the nVidia sites react to the whole DoomIII & R300 announcement.

Do they have to react to it at all? No one reacted to GF3/DoomIII announcment at Macworld, at least I don't recall that.

As for R300 being further along the line then NV30 - I think that is to be expected considering that while ATi moved directly from R200 to R300 Nvidia has released NV25 between NV20 and NV30. Even with parallel development, a sizeable commitment of the company recourses to the NV25/XGPU project is bound to detract from NV30 development.

 

[Joe DeFuria]

Do they have to react to it at all?

No, they don't really. That's why it's fun watching them react. For example, there's a web-master at another site raising questions like "This does make me wonder, did ATI clean up the drivers for the R300 only, and are neglecting the R8500 then? "

ATI just can't be "congratulated" for this "win" and have it left at that...let's attach some strings.

As for R300 being further along the line then NV30 - I think that is to be expected considering that while ATi moved directly from R200 to R300 Nvidia has released NV25 between NV20 and NV30.

That doesn't make much sense to me. ATI will have moved from the R200 to the R300 in a little less than a year. nVidia has had about 18 months to move from the NV2x core to the NV30.

 

[pascal]

What can we expect?

- R300 will be available before the NV30.
- The drivers will have more time to mature and will be Doom3 ready.
- Performance will be good.
- Some good prices by the end of the year.

More competition in the market

 

[Colourless]

I think it is more useful to compare the times between major revisions, not just the last 2 chips for each company

So for Nvidia its
NV10 (1999)->NV20(2001)->NV30(late 2002??)

and for ATI
R100 (2000)->R200 (2001)->R300(mid 2002??)

It looks to me that ATI has been managing to work somewhat quicker than Nvidia if you think about it.

R100 came out after NV15.
R200 came out about half way between NV20 and NV25.
R300 is likely to come out just before NV30.

At the moment it seems that ATI might almost be making major revisions as fast as Nvidia can make a refresh.

-Colourless

 

[muted]

well here's a press release

and it's coming this fall

http://www.businesswire.com/cgi-bin/f_headline.cgi?day0/221442087&ticker=atyt|aty.to

 

[maguen]

At the moment it seems that ATI might almost be making major revisions as fast as Nvidia can make a refresh.

-Colourless

From a hardware perspective, I agree. But product release timing consists of software, hardware and market considerations.

I remember last year when there was talk about working NV25 hardware being shown months before it was released. I believe that Nvidia seems to take longer getting the software ready for each revision, probably because their driver regression/process is very long/strict (A guess). After all, driver quality has to come from somewhere, and that development no doubt takes time.

How would ATI look in your comparison if they had waited to the current or more mature driver set was available, before releasing the hardware ...

They (Nvidia) also seem to purposely delay products to satisfy OEM concerns about saturation. The delay of the Ti4200 looks quite intentional ...

One thing is for sure though, the next round will be very interesting to watch ...

Guenther

 

[Colourless]

maguen true. NV20 was also rumoured to have been in silicon for quite a while before it was even announced.

 

[psurge]

Well it looks like r300 is dx9.0. So do people think nv30 is dx9.1 (i am speculating based on the "revolutionary" new architecture rumors)?

The wait is killing me... anyway, whatever. DaveB when do you think you will have the p10 interview posted?

 

[Dave Baumann]

Late next week / early week after.

 

[Hellbinder]

WEll this puts this discussion to rest. As well as all this talk about Ati having their drivers ready for doomIII *by the time it ships*....

Press Release from ATi today.

E3 2002, Los Angeles, CA - May 24, 2002 - Id Softwareâ„¢ has chosen ATI's next generation graphics architecture as the platform for previewing their unbelievably realistic and much anticipated DOOM IIIâ„¢ at this year's E3 Expo.
ATI's chip was chosen because it offers the most advanced, high-performance graphics architecture to power the demonstration at E3 2002. DOOM IIIâ„¢, scheduled for release in 2003, has been highly anticipated both as an unbelievable game and as a true measure of today - and tomorrow's - graphics architectures. Also highly anticipated, ATI's next generation product, scheduled for release this fall, will cement ATI firmly as a leader in the ultra-high-end 3D and gaming graphics market.

"ATI's next generation hardware has an ideal feature set for the DOOM IIIâ„¢ engine, and at the moment is the fastest platform to run the game on," said John Carmack, Chief Technical Officer & co-founder of Id Software, Inc.

"We want users to be blown away by their gaming experience," said Rick Bergman, Senior Vice President, Marketing and General Manager, Desktop, ATI Technologies Inc. "Our next-generation product will move gamers a quantum leap forward in realistic and immersive gaming. Gamer's won't just play DOOM IIIâ„¢ with ATI's next generation hardware - they'll live it."

This speaks for iteself... especially since JC also has access to Nv30..