View Full Version : nVidia CEO: NV30 taped out/120M transistors
Typedef Enum
06-Jun-2002, 06:48
I don't have the magazine, but one of the guys @ 3dgpu said that the latest issue of Wired magazine features an interview with nVidia's CEO, and he states that NV30 has, indeed, taped out and will contain approx. 120 Million transistors.
http://www.3dgpu.com/yabb_se/index.php?board=2;action=display;threadid=525
Taking this info @ face-value...I think it pretty much lays to rest the notion of a second chip...With a chip of that size, it seems very difficult to imagine a co-processor consuming even more real estate.
Maybe all that additional logic is for that Glide compatability we keep hearing about... :)
:o :o :o
Holy crap! I have nothing useful to post, other than it seems like a monster.
Wow. So, are we guessing dual/quad pixel shaders? Perhaps more vertex shaders than R300's and Parhelia's four?
A doubling of NV25's die seems to be no small upgrade.
what tha hell they are building?? The Holy HandGrenade??
(check the Monty Python's movie 'Quest for holy grail.') :o
mboeller
06-Jun-2002, 08:15
How fast will it be?
120Mio Transistors @0,13µm <-> 63Mio @ 0,15µm.
Maybe it will only work at 250MHz. Or are most additional transistors for eDRAM?
Typedef Enum
06-Jun-2002, 08:28
How does that cliche go?
Expect the unexpected? That's what I'm thinking...Honestly, nVidia has had such a predictable history, it becomes really difficult to believe that they would veer off course, so to speak...But I think the realization that the competition was slowly catching up to her may have had some impact in this one.
I haven't read Wired to verify the authenticity of this claim...But assuming it's true, it just seems unthinkable that they could double the transistor count from a GeForce4...But if the CEO said it...it seems rather credible.
Randell
06-Jun-2002, 09:03
well they have hinted at a 'discontinuity' in their next chip havent they?
PVR_Extremist
06-Jun-2002, 09:16
So its gonna be a sound card :lol:
First announcement (remember I said sooner than later :D) Now the wait for the 2nd shoe to drop on NV30 ;) . Soon :)
Evildeus
06-Jun-2002, 09:43
First announcement (remember I said sooner than later :D) Now the wait for the 2nd shoe to drop on NV30 ;) . Soon :)You are the devil :evil: Ben6 ;)
This fall will really be interesting!
Chalnoth
06-Jun-2002, 10:27
Heh, Glide, riiiight....that'd be about 40 steps backward...and I think it would make everybody else in the industry very, very angry at nVidia (we really don't need another proprietary API). There is just absolutely no possibility that we'll see Glide ressurected.
But personally, I fully believe the 120 million transistor number. nVidia has been saying that this will be the most significant contribution to 3D graphics they've ever made. The big question, then, is why? Will the NV30 have much more flexible programmability (similar to the P10)? Or will it be more on the software side, with more accessible programmability?
Of course, we can be about 99.9999999% certain that the NV30 will have full 16-bits per color floating-point pipelines. But, this is really no more significant toward advancing 3D technology than 32-bit color was. And the usage of just more advanced pixel and vertex shaders wouldn't be any more significant than the initial introduction of pixel and vertex shaders.
So, why is this NV30 to be so significant? I'm really excited about finding out :)
Gunhead
06-Jun-2002, 10:42
Maybe something in the L part of TCL? 8)
...that this will be the most significant contribution to 3D graphics they've ever made.
I have no doubts about this, but I what I fear is huge amount of very advanced features that no one is willing to support yet, because it would be make game developing go back to direction where it was when proprietary APIs were only way to go. (regular version for "normal" cards and special version for NV30.) and I agree with Chalnoth that this is not a nowadays way to do this.
Cards that are too much ahead of their time, never gets fully supportted their feature sets at least with playable speeds.
for example, Pyramid3D was a way to much ahead of it's time. even if it would have stayed on schedule, It's features would have been fully supportted not until DX8 and for that it would have been a way too slow.
(think of it... card with equal memory bandwidth of TNT Vanta and with full DX8 feature set.... huh?)
another side of the coin is, somebody needs to take a step, eventhough it would mean feature set that gets no support at all during chip's lifetime.
Dave Baumann
06-Jun-2002, 10:58
And we don't necessaily know that, on chips these sizes, the supposed extra 13million transitors over the supposed 107 of R300 will actually necessarily bring much to the table in terms of features...
But think of it... What the heck could Nvidia use 120M transistors for?
eDRAM of some sort seems a bit unlikely in my opinion, they've never even hinted at being interested in that direction before. And 16-bit per color component FP pipes hardly consumes THAT much more either.
So are they simply heaping on (much) more of the same, like 8 pixel/vertex shaders etc etc? Or what?
I saw a list of possible topics for some Nvidia conference to be held around siggraph. I read stuff like solid modelling and non-graphics use of GPUs and stuff. Maybe that's what NV30 uses all those extra transistors for.
*G*
Brimstone
06-Jun-2002, 11:28
I read some time back that it will be possible for 2 AGP 8x slots to function on a motherboard. I have no idea if this is true, but if it is the case I could see NVIDIA allowing for SLI on their new cards.
Does anyone know if the AGP 8x spec allows for two AGP slots?
AGP 8X (in AGP 3.0) supports a fan-out bridge for two AGP cards sharing a AGP bus. Of course, the motherboard manufacturer have to implement it.
Or maybe you can put it on the video card too? (Obviously it would make 2 chip setups a lot easier.)
Ugh... I am not sure about this. IIRC there can be only one fan-out bridge on a AGP bus in AGP 3.0. If there is a fan-out bridge on a video card, what happens when it is put on a motherboard with a fan-out bridge?
Or maybe we need a sticker on the video card says "For MBs with one AGP slot only" :)
From TSMC process info NV30 with 120millions transistors the die area should be equal to a 90millions transistors .15 micron chip :o
This chip will not be cheap
If it has only 128bits bus then we can expect some big texture caches.
From TSMC process info NV30 with 120millions transistors the die area should be equal to a 90millions transistors .15 micron chip :o
This chip will not be cheap
If it has only 128bits bus then we can expect some big texture caches.
If a 90mil .15 micron chip is so bad, where does that put the 107mil transistor R300 at .15um considering cost?
I really dont know, but .15 micron now should have better yields than the .13 micron process (for now).
Also this 120millions come from the wired magazine nvidia CEO interview. Where did this 107 come from?
What I am asking myself is when we will have a sub $200 DX9 card? :-?
Pascal: I guess that SiS will be showing Sub $200 DX9 card first...
In their roadmap it was marked to Q1/2003
Ailuros
06-Jun-2002, 15:10
Pascal: I guess that SiS will be showing Sub $200 DX9 card first...
In their roadmap it was marked to Q1/2003
I have something far more capable in mind and with far more solid drivers ;)
Evildeus
06-Jun-2002, 15:43
Some more info:
http://www.tweaktown.com/document.php?dType=review&dId=279&dPage=7
Derek Perez told us that we can look forward to film-like image quality with the NV30 and if this is true, there are certainly gonna be some exciting times ahead on the gaming front.
Chalnoth
06-Jun-2002, 16:02
Pascal: I guess that SiS will be showing Sub $200 DX9 card first...
In their roadmap it was marked to Q1/2003
I don't believe we'll see a full 64-bit color-capable DX9 part by then at under $200. I'd say that's probably going to be about a year from now, actually...
But what this is really shaping up to be is that everything else out there is going to be more of a DX8.5 part (Including the R300), and the NV30 is going to be the only true DX9 part...
Doomtrooper
06-Jun-2002, 16:07
But what this is really shaping up to be is that everything else out there is going to be more of a DX8.5 part (Including the R300), and the NV30 is going to be the only true DX9 part...
Ummmm No... :roll:
Dave Baumann
06-Jun-2002, 16:07
I don't believe we'll see a full 64-bit color-capable DX9 part by then at under $200. I'd say that's probably going to be about a year from now, actually...
But what this is really shaping up to be is that everything else out there is going to be more of a DX8.5 part (Including the R300), and the NV30 is going to be the only true DX9 part...
I think its a bit premature to be concluding that for R300 just yet, we are still not aware of all its features. BTW, I'm not sure if you were implying it or not but AFAIK 64bit isn't required for DX9, I believe 10:10:10:2 is the DX9 requirement.
Joe DeFuria
06-Jun-2002, 16:10
But what this is really shaping up to be is that everything else out there is going to be more of a DX8.5 part (Including the R300), and the NV30 is going to be the only true DX9 part...
I would also like to know how you came to that conclusion, specifically, concerning the R300.
Chalnoth
06-Jun-2002, 16:17
I think its a bit premature to be concluding that for R300 just yet, we are still not aware of all its features. BTW, I'm not sure if you were implying it or not but AFAIK 64bit isn't required for DX9, I believe 10:10:10:2 is the DX9 requirement.
Nothing's required for DX9-compatibility. It's just that it's really shaping up like there will be that much of a difference between the NV30 and everything else. Of course, I don't have any real information to back this up, it just looks like this is the shape things are going to take...
And I REALLY hope that 10:10:10:2 is NOT what ATI is doing, as that is one step forward and two steps back. Games really start to need using that alpha channel for improved visual quality, and the widespread usage of that sort of color depth would push that back another 3-5 years...
dksuiko
06-Jun-2002, 16:21
Nappe1,
"It would be make game developing go back to direction where it was when proprietary APIs were only way to go. (regular version for "normal" cards and special version for NV30)"
Believe it or not, I would love to see that again. If any graphics company, whether it be nVidia or ATI or someone else, put out a graphics card with a proprietary API and game developers actually USE those features in making SPECIAL versions for those cards - that wouldn't be a bad thing, that'd be great! New graphics cards have almost become synonymous with: "YAY! More useless features that will never to be used in games!" I'm all for the progress of new technology and it's fun to see what they can do in demos, but I think it'd be even better to actually put all of those features to use in a game.
If a proprietary API can change the slow adoption of new features - I'm all for it, but I doubt they will make a difference at all when it comes to that. One of the reasons that proprietary APIs like Glide were successful was because game developers wanted to USE the new features not available in APIs like DirectX. Now, they either don't want, don't feel the need, or can't (consumer base, minium requirements, or whatever) use the new features in the latest video cards. Anyway, if the release of a proprietary API would encourage developers to make SPECIAL versions for the new cards, then I'm all for it. But then again, they aren't using the latest features NOW in DirectX8 - new features in a future proprietary APIs won't be any different.
-dksuiko
Colourless
06-Jun-2002, 16:30
Well, why not just use the vendor specific OpenGL extensions?
Ailuros
06-Jun-2002, 16:31
Well if rumours are true that the two big boys are again racing against each other wether MS will include this or that extension, then I could assume that they are pretty close in compliance.
I wouldn't be surprised either if the scenario would be the other way around. In the end is it really that important? The bottomline is which card will be faster with the games that are or will be soon on shelves. And in that aspect I don't see any dx9 games coming that soon after all.
As far as the sacrificed bits in the alpha channel concerns then Matrox and 3DLabs seem to be "out of their minds too"..... :lol:
dksuiko - companies can expose any new feature they want by releasing a proprietary OpenGL extension. That gives them 99% of the benefit of providing their own API, with the aditional benefits of increased developer support, ease of implementation, and the shield of political correctness.
NV30 Features - I doubt the NV30 contains any significant API-level features we don't already know about. I predict that every important feature of the NV30 is exposed in DX9.
It just wouldn't make sense for NVIDIA to cripple their chip by not getting support for their chip's major features into DX9.
The DX9 beta is out, and so far I haven't heard of any groundbreaking new features beyond what we already knew. Because of this, I predict that the NV30 will be more of a refinement of the NV25 than a radical new architecture.
I expect more programmability, more parallelism, and higher performance, but that's about it.
[/b]
arjan de lumens
06-Jun-2002, 16:43
Hmm .. what kind of effects might the alpha channel in the framebuffer be useful for, other than overlays (which really only require 1 bit)?
Well, with a high precision alpha buffer (8 bits is a bit too limited) you could use it for depth in volume calculations for volumetric effects (fog, or light-beams through dust type stuff). But if the hardware supported a single channel high precision format then it could use an offscreen buffer just as well of course.
Chalnoth
06-Jun-2002, 16:54
NV30 Features - I doubt the NV30 contains any significant API-level features we don't already know about. I predict that every important feature of the NV30 is exposed in DX9.[/b]
I just have to say, what exactly are those features? I do know about 64-bit color, and in general that there will be better HOS support, but little else...
Reverend
06-Jun-2002, 17:31
Grrr, don't you now just hate all those websites who already have been NDA'ed on NV30....
Nothing's required for DX9-compatibility. It's just that it's really shaping up like there will be that much of a difference between the NV30 and everything else.
Again not to be a butthead but I would really like to know where this is coming from? Normally you have been very objective and its not like you to make such claims. Thats why I am wondering if you have found some other creditible information.
I personally dont think that the NV30 will have that many supper great improvements of the R300 at the time of the lunch. Sure it will have lots of neat things but both parts will be more or less DX9 compliant.
What I hope does not happen is that we get the same fraction that we did with PS1.0 1.1, 1.3 & 1.4 Granted they all probably had their reasons...but it did get kind of confusing at times...
But what this is really shaping up to be is that everything else out there is going to be more of a DX8.5 part (Including the R300), and the NV30 is going to be the only true DX9 part...
Very unlikely. It could be that even NV30 does not support all new DX9 features. Personaly, I seriously doubt that NV30 will support the Matrox kind of Displacement Mapping.
Joe DeFuria
06-Jun-2002, 18:33
Again not to be a butthead but I would really like to know where this is coming from?
I would like to know as well. I don't see any indication that NV30 should be any more significantly "advanced" than R-300.
120 million vs. 107 million transistors?
I do find it interesting that (apparently) nVidia has already "briefed" web sites on NV30 features under NDA, and yet they have just "taped out." On the other hand, I don't beleive ANYONE (except for ATI's board partners) has been briefed on R-300. Anand certainly hasn't...ATI flew him out for a visit last week...and they told him nothing of R-300. ATI is really keeping it under wraps, even though they obviously have working samples.
I do know about 64-bit color, and in general that there will be better HOS support, but little else...
What about 64 bit color? Does that mean 64 bit FRAME buffer? Or 64 bit internal calc precision?
Anyway, I, like most, would wager that the relative "power" (performance and features) of the NV30 and R-300 will probably be pretty close. And in reality, they will each probably have their relative advantages and disadvantages.
Hm.
There's a short interview with Dave Kirk of Nvidia in the recent issue of the German games magazine PC GAMES.
Very rough translation:
PC GAMES: "What's your opinion on doubling memory bandwidth as Matrox did with Parhelia?"
DAVE KIRK: "I like the 'brute force' approach. The Riva128 was the first chip to implement a 128bit wide bus and it was successful. Doubling bus width is a question of costs though, since you also need to extend the 3D pipeline in order to fully utilize it."
He also said that Doom3 will run smooth on GF4Ti cards but if you want to play in "high resolutions" (whatever that means) or use AA you need a next generation part.
Daemon_UK
06-Jun-2002, 20:13
[quote]Personaly, I seriously doubt that NV30 will support the Matrox kind of Displacement Mapping.
Isn't Displacement Mapping part of the DX9 spec? I could be wrong tho.
Geeforcer
06-Jun-2002, 20:18
I do find it interesting that (apparently) nVidia has already "briefed" web sites on NV30 features under NDA, and yet they have just "taped out." On the other hand, I don't beleive ANYONE (except for ATI's board partners) has been briefed on R-300. Anand certainly hasn't...ATI flew him out for a visit last week...and they told him nothing of R-300. ATI is really keeping it under wraps, even though they obviously have working samples.
First off, "just"? What is the lead time of Wired? Second, I would be surprised to learn that Anand in not briefed and NDA'ed about R300.
Joe DeFuria
06-Jun-2002, 20:25
http://www.anandtech.com/showdoc.html?i=1632
Just a few weeks ago we were in Toronto visiting ATI and they were very tight lipped about anything R300 related; it will be interesting to see if VIA was supposed to be publicly running this R300 in their suite....Given that the chip is already up and running and production is due soon we are beginning to wonder if the R300 will be made on a 0.13-micron process or if it will be 0.15-micron like its predecessor. If it is indeed a 0.15-micron chip then there is the question of whether ATI will make it a DX9 compliant part with full floating point pipelines. Assuming ATI does make the R300 as feature rich as NVIDIA's NV30 currently appears on paper, then there's the question of yield and clock speeds. It will be interesting to see the design choices ATI made with the R300 and how that effects competition with the NV30 later this year.
Doesn't sound to me like Anand has any official briefing on R-300. So what was anand doing in Toronto?
http://www.anandtech.com/showdoc.html?i=1633&p=2
We've already been briefed under NDA on ATI's RV250 part so we cannot share the majority of material about the RV250 chip but we can definitely offer what we've learned today.
In other words, because Anand is under NDA for RV250, he cannot really talk about RV250 specifications at all. It is only because he is NOT under NDA for R-300, that he can freely speculate on specs and talk about what he saw at VIA's stand.
Geeforcer
06-Jun-2002, 20:33
Just because he is under NDA doesn't mean he can't discuss information obtained by other means. Besides, it is not known how inclusive ATi NDA is. That said Assuming ATI does make the R300 as feature rich as NVIDIA's NV30 currently appears on paper... does indeed seem to indicate that he is NDA'ed for NV30 but knows nothing of R300.
LeStoffer
06-Jun-2002, 21:13
First announcement (remember I said sooner than later :D) Now the wait for the 2nd shoe to drop on NV30 ;) . Soon :)
So I take that the 120 million transistor number is true, ben6? (I cannot find anything about the NV30 in the Wired magazine (June) on www.wired.com)
Chalnoth
06-Jun-2002, 21:29
Again not to be a butthead but I would really like to know where this is coming from? Normally you have been very objective and its not like you to make such claims. Thats why I am wondering if you have found some other creditible information.
I have heard little more than whispers. This is more just a personal belief of what will happen than a statement of any knowledge of insider info. It's mostly based on past history.
Controls-techy
06-Jun-2002, 23:51
I agree with some of the other forum members that eDRAM is unlikely with this chip design. The reason being is the more than likely 256 bit, ~700 MHz DDR interface to external RAM. I mean why bother with eDRAM when you have ~22GB/sec bandwidth?
Cheers.
Oompa Loompa
07-Jun-2002, 00:01
I do find it interesting that (apparently) nVidia has already "briefed" web sites on NV30 features under NDA, and yet they have just "taped out".
We don't really know when it taped out. During the Q1 conference call they spoke about NV30 and about their "record number of new chip tapeouts". They've never specifically said when it taped out. Timing-wise all we know is it'll supposedly be launched in August, with Fall availability (again as per the Q1 CC).
It is strange that ATI hasn't been more forthcoming. It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention. Their only high margin part is the low volume AIW.
Joe DeFuria
07-Jun-2002, 00:47
It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention.
On the contrary. The Radeon 8500 chip, AFAIK, is a great seller for ATI. I'm sure ATI's board partners want as the R-300 product launch to be as close as possible to actual shipping. Otherwise, they can be left holding lots of Radeon 8500 stock, and forced to make price cuts to unload them.
DemoCoder
07-Jun-2002, 02:45
What do we really know from the publically available facts on the NV30 and R300?
Carmack: R300 beat suped up GF4 in every test. (didn't say by how much?)
If by 1.2x, ATI is in trouble, if by 2x, they might hold against the NV30 if we assume NV30 = 2x-3x NV25
Wired: 120mil transistors, .13um for NV30, 107mil/.15um for R300
.13um could translate into higher clock, lower heat and power and even higher yields and lower costs for NVidia. On the other hand, it is a new process and could introduce bad yields and significant delays.
Various: DX9 is modelled more after ATI.
ATI has been driving DX and OGL more lately. Displacement mapping seems to be driven by Matrox however. ATI and NVidia could BOTH lack displacement mapping. NV30 could lack DX9 features that R300 has and vice versa. We have no information, but I would expect the R300 to be more compliant since ATI is driving it. NV30 could have as many features as the R300, but they will be different NV30 specific features that NVidia will have to expose through OpenGL extensions probably. NV30 and R300 might have DX9 subset overlap, but both might have "beyond DX9" features.
3dfx: Bought Gigapixel, bought by NVidia. NVidia also employs Ned Greene. NVidia now owns some of the top researchers and engineers (and patents) in the field of hardware visibility. NV30 could implement Gigapixel tech, hierarchical-z based occlusion culling (Greene's approach), or some hybrid. ATI has ArtX team. Reusing any Flipper tech?
R300 is implementing Hyper-Z III which sounds evolutionary, not a significant departure or completely new algorithm.
Nvidia exec: NV30 is a sea change, biggest thing they did since 3D graphics started. First non-GeForce/non-TNT pipeline architecture?
I'll tell ya, the suspense is killing me. I'm excited by the coming lineup: Parhelia, P10, NV30, and R300. When have we had 4 new chipsets with significant new feature sets show up in such a short time period?
A year ago, it looked like everyone else was dead, and this was going to be a two-horse race between NVidia and ATI like Intel and AMD.
Hey Ben,
See, I'm not the only one to know Ned Greene. Very, very smart man - does he still go to SIGGRAH?
It is strange that ATI hasn't been more forthcoming. It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention. Their only high margin part is the low volume AIW.
Naw your forgeting the lion share of the $$ to be made is at the lower end scale were you have more sales. And they are not doing too bad there.
Chalnoth
07-Jun-2002, 03:37
Various: DX9 is modelled more after ATI.
I seriously doubt that this is true. Why would Microsoft design their API more for the company that has less influence on the market as a whole? It just doesn't make any sense.
Joe DeFuria
07-Jun-2002, 03:55
Why would Microsoft design their API more for the company that has less influence on the market as a whole?
Yes, Chalnoth, we know that you seriously doubt anything could be true that would favor ATI over nVidia. :wink:
Why don't you think up an answer to your own question. Just ask yourself why would Microsoft bother to release Pixel Shader 1.4 support for their API, if nVidia has (so you claim) so much more influence on the market? :roll:
Chalnoth
07-Jun-2002, 04:02
Why don't you think up an answer to your own question. Just ask yourself why would Microsoft bother to release Pixel Shader 1.4 support for their API, if nVidia has (so you claim) so much more influence on the market? :roll:
Why wouldn't Microsoft add support for ATI hardware? Microsoft has always been very forward with adding support for every feature that hardware designers have come up with.
Joe DeFuria
07-Jun-2002, 04:12
Well, then why would you "seriously doubt", considering that R-300 will be here before NV-30, that DX9 would be "more modeled" after ATI hardware?
You are just being terribly contradictory in your logic IMO.
Chalnoth
07-Jun-2002, 04:22
Well, then why would you "seriously doubt", considering that R-300 will be here before NV-30, that DX9 would be "more modeled" after ATI hardware?
Why will DX9 be more modeled after ATI hardware? I seriously doubt that it will pan out that way. I think we'll find out that, in the end, DX9 will fully support both the NV30 and R300, and it will be very hard to determine which was DX9 was modeled more after. The only people that will know for sure as those involved in the decisions at MS, ATI, and nVidia.
But I still think that the NV30 will have significantly more features, and thus be considered the only "true" DX9 part...
Joe DeFuria
07-Jun-2002, 04:26
Why was DX 8.0 more modeled after nVidia hardware...
Chalnoth
07-Jun-2002, 04:34
Because the GeForce3 design was finalized long before the Radeon 8500's (which is different from the release of these cards). nVidia also pioneered in the original pixel shader tech, which is currently defined as what makes a DX8 card a DX8 card.
Doomtrooper
07-Jun-2002, 04:40
Well, then why would you "seriously doubt", considering that R-300 will be here before NV-30, that DX9 would be "more modeled" after ATI hardware?
Why will DX9 be more modeled after ATI hardware? I seriously doubt that it will pan out that way. I think we'll find out that, in the end, DX9 will fully support both the NV30 and R300, and it will be very hard to determine which was DX9 was modeled more after. The only people that will know for sure as those involved in the decisions at MS, ATI, and nVidia.
But I still think that the NV30 will have significantly more features, and thus be considered the only "true" DX9 part...
http://www.subzerotech.com/articles/interviews/ati/pgs/pg1.shtml
Q32: Is there a chance the next generation video cards following Radeon 8500 would support 8X standard and DirectX 9.0?
A32: RADEON 8500 is the first graphics technology to support the DirectX 8.1 specification. Through its advanced SMARTSHADER technology that takes advantage of Microsoft’s new DirectX 8.1 specification, including supporting Pixel Shaders Version 1.4, it enables more complex and realistic texture and lighting effects without sacrificing performance. DirectX 8.1 is an integral part of Microsoft’s new Windows XP operating system. RADEON 8500 will be the reference platform for Microsoft’s development of DirectX 9.0.
I guess you're wrong :)
I seriously doubt that this is true. Why would Microsoft design their API more for the company that has less influence on the market as a whole? It just doesn't make any sense.
Yes it does, it's all 'political'. MS DX got burned by nVidia when they [nVidia] forced MS in the 11th hour to liecense the core fragment shading technology/implimentation for use in DX8.
Microsoft was pist and shortly thereafter it became known that the DX group switched to ATIs R200 for DX development and used the R300 for the target reference platform. Thus, due to the fact that MS is/was/suppose to design DX9 toward the R300s features (Just like NV20 was the target for DX8.0) that it would have a better/more complete DX9 featureset.
Although, With a transistor count of 120M - it's appearent that these transistors are going toward programmability (as the two are directly related; transistor count - programmability) in the pipeline. Thus, the supported 'features' becomes blurrier and blurrier as the pipe becomes more programmable. NV30 will most likely be DX9 complient either threw design intention or it's inherient flexibility.
Anyways, the DX9 thing is True story :) I thought this was well know.
DemoCoder
07-Jun-2002, 04:58
Chalnoth,
People here know that I am by no means anti-Nvidia. Hell, most would consider me an NVidia apologist. Take a look at Pixel Shaders 1.4, now take a look at Pixel Shaders 2.0. The architecture of PS2.0 is an extension of PS1.4. Clearly, the ATI's architecture has influenced DX9 PS. Many of the new PS2.0 instructions and registers have the same names as their PS1.4 counterparts, whereas many of the old tex* DX8 PS1.0 instructions have been deprecated.
In regards to pixel shading, it seems DX9 is based on ATI's texturing and combiner pipeline, not Nvidia's.
As far as political/PR skill goes, ATI has successfully outmaneuvered Nvidia within both Microsoft and OpenGL. Microsoft likes to keep the market fragmented. If a single manufacturer gets too powerful, they can start dictating to Microsoft. MS's preffered situation would be to keep both ATI and NVidia out of monopoly position. Ironic, isn't it?
Oompa Loompa
07-Jun-2002, 05:38
It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention.
On the contrary. The Radeon 8500 chip, AFAIK, is a great seller for ATI.
Volume is good, margins are not. Rather like AMD, actually.
I'm sure ATI's board partners want as the R-300 product launch to be as close as possible to actual shipping. Otherwise, they can be left holding lots of Radeon 8500 stock, and forced to make price cuts to unload them.
I think it's safe to say that all of the next generation products will debut at a much higher price point than the R8500, which is going for $95 on the street. Consumers considering the purchase of a ~$100 card will not be swayed by rumors of a $300-$400 card.
NVIDIA's customers, on the other hand, already have a three-month-old product at that approximate price point. It is their sales that would be cannibalized by aggressive discussion of next generation hardware from a credible manufacturer (ie. NVIDIA or ATI). Creative and Matrox can talk all they want - at this point, no one is listening.
nVidia also pioneered in the original pixel shader tech, which is currently defined as what makes a DX8 card a DX8 card.
umm... anyone remembers PS 1.0 / VS 1.0 ??
they were for DX8 and made by 3dfx, if I am not mistaken...
How about nVidia's NSR and ATi's Radeon programmable pixel pipeline?
how about Pyramid 3D?
and what comes to general programmability, how about Rendition Verite chips with microcode programmable core? or even PixelFusion's PixelFuzion 150 chip?
I don't think nVidia invented pixel shader tech... it has existed a much longer.
Chalnoth
07-Jun-2002, 06:56
Chalnoth,
People here know that I am by no means anti-Nvidia. Hell, most would consider me an NVidia apologist. Take a look at Pixel Shaders 1.4, now take a look at Pixel Shaders 2.0. The architecture of PS2.0 is an extension of PS1.4. Clearly, the ATI's architecture has influenced DX9 PS. Many of the new PS2.0 instructions and registers have the same names as their PS1.4 counterparts, whereas many of the old tex* DX8 PS1.0 instructions have been deprecated.
Guess what, dude? I can't look at PS 2.0. Where's your info?
A32: RADEON 8500 is the first graphics technology to support the DirectX 8.1 specification. Through its advanced SMARTSHADER technology that takes advantage of Microsoft’s new DirectX 8.1 specification, including supporting Pixel Shaders Version 1.4, it enables more complex and realistic texture and lighting effects without sacrificing performance. DirectX 8.1 is an integral part of Microsoft’s new Windows XP operating system. RADEON 8500 will be the reference platform for Microsoft’s development of DirectX 9.0.
Yes, I remember reading that a long time ago, and I feel it's just chock-full of PR bull. It's just close enough to correct to not be false advertising :P (not that nVidia does anything different...)
Dave Baumann
07-Jun-2002, 09:20
Guess what, dude? I can't look at PS 2.0. Where's your info?
The Microsoft Meltdown slides have had information on PS2.0 for a long time.
http://www.microsoft.com/mscorp/corpevents/meltdown2001/ppt/DXG9.ppt
What do we really know from the publically available facts on the NV30 and R300?
Carmack: R300 beat suped up GF4 in every test. (didn't say by how much?) Carmack has confirmed that that "high clocked" GF4 was in fact a retail 4600. If we're talking "Free FSAA." then we're looking at not-insignificant increase. I'm not sure how much memory bandwidth affects FSAA, but 20GB/s couldn't have hurt.
Dave Baumann
07-Jun-2002, 10:22
I'm not sure how much memory bandwidth affects FSAA, but 20GB/s couldn't have hurt.
If its multisampling then dependant on how your pipes are arranged bandwidth can more or less be the only limiting factor.
umm... anyone remembers PS 1.0 / VS 1.0 ??
they were for DX8 and made by 3dfx, if I am not mistaken...
AFAIR the (original) Radeon had PS1.0 which should have been exposed in DX8.0, but shortly before the DX release, MS changed the minimum PS in DX8.0 to version 1.1(nV), thus rendering the Radeon PS more or less useless...
Dave Baumann
07-Jun-2002, 10:58
Radeon falls short of even PS1.0 specification.
Oh well, then I didn't remember correctly... ;)
But there WAS something regarding Radeon and changing PS specs, or am I completely wrong? Someone refresh my memory, please? :)
Randell
07-Jun-2002, 11:09
at its launch ATI claimed the Radeon was 'more' DX8 compliant than the Gf2 and V5. The original Pixel Tapestry engine did have some shading ability, like the NSR did in the Gf2.
They never cliamed DX8 compliance though.
Oh well, then I didn't remember correctly... ;)
But there WAS something regarding Radeon and changing PS specs, or am I completely wrong? Someone refresh my memory, please? :)
yep, there was... as said in the earlier post (in mine as also in Randell's.), Pixel Tapestry engine had Programmable pixel pipeline, but it was decided to be controllable via DXTextureStages functions and because of that it became more or less useless. (somewhere it was said that Pixel Tapestry could said be PS 0.5)
But whose implemtation was that PS 1.0 which is a part of DX8?
Only publically shown chip with programmable pixel pipeline I can remember is Pyramid3D... if PS 1.0 needs only two textures on single pass, it could be Legacy of Tritech... (I weakly recal that PS 1.0 definition would have been finalised much before DX8 and that it would be made with some company that doesn't exists anymore. So it could be also 3dfx.)
And how about VS 1.0? it exists but what are the main differecies to VS 1.1?
Does anyone know the exact answers?
Joe DeFuria
07-Jun-2002, 13:32
And this is why it's quite maddening to try and have a debate with you, Chalnoth: you say things like this:
Because the GeForce3 design was finalized long before the Radeon 8500's (which is different from the release of these cards).
as if this is a fact, without showing any evidence. Where is it written that NV-20 design was "finalized" long before R-200?
And with R-300 supposedly on a 0.15, and NV-35 supposedly on 0.13, which of THOSE designs do you suppose was "finalized" first?
nVidia also pioneered in the original pixel shader tech, which is currently defined as what makes a DX8 card a DX8 card.
Another completely unfounded statement. Just becuase nVidia had produced the first hardware that had complied with DX8, doesn't mean anything. Most known info would have ATI, nVidia, and 3dfx all developing "shaders tech" in parallel. 3dfx's "texture computer" which was apparently PS 1.0, nvidia's shader tech (1.1-1.3) , and ATI's shader tech, 1.4.
Reverend
07-Jun-2002, 15:03
As much as I know Chalnoth knows what he's talking about in certain subjects, he's becoming obvious in his alliance with a certain company. As well as displaying his lack of experience with cards other NVIDIA-based, which in turn tells us whether his comments vis-a-vis comparisons are valid or not.
MDolenc
07-Jun-2002, 15:36
Come on guys get serious...
There is actually no (at least I'm not aware of any) hardware that supports PS.1.0 (because all hardware that supports PS.1.0 also supports higher versions). But it is very hard to say which company "invented" pixel shaders, because all pixel pipelines since Voodoo 2 are highly configurable. I'd say that it was actually Microsoft's idea, because pixel shaders were meant to introduce standard for multiple hardware vendors (DirectX for X-Box for example has much different pixel shaders (register combiner like) even if the hardware is almost identical to GeForce 3).
All 1.x pixel shaders can have up to 8 arithmetic instructions (PS.1.4 has two phases and can do 16 instructions through loop back). Pixel shaders 1.4 can use up to 6 different textures where all previous versions can use only 4. There are only small differences between PS.1.0 and PS.1.1 (texbem and texbeml count as two address instructions in PS.1.0 and one in PS.1.1, PS.1.1 can use different registers for output when pairing two instructions). PS.1.2, 1.3 have two arithmetic instructions and few texture addressing instructions more than PS.1.0 and PS.1.1. PS.1.4 has one arithmetic instruction more then PS.1.3 and only 4 texture addressing instructions. Texture addressing and arithmetic instructions can be mixed in PS.1.4.
PS.2.0 is of course way ahead with 64 arithmetic instructions and 32 texture addressing instructions and again two phases...
But I'd say that vertex shaders were done (almost?) completely by NVidia (OpenGL extensions behave exactly the same as VS.1.1). The only difference between VS.1.0 and VS.1.1 is that VS.1.1 has address register and VS.1.0 does not.
There was actually so called PS.0.5 in DirectX 8 Beta that Radeon and GeForce 1/2 supports, but it was dropped before DirectX 8 made it public. NVidia had drivers that supported this version but dropped support when Microsoft decided that is does not want PS.0.5, but ATI renamed it into 1.0 but again dropped support latter.
People here know that I am by no means anti-Nvidia. Hell, most would consider me an NVidia apologist. Take a look at Pixel Shaders 1.4, now take a look at Pixel Shaders 2.0. The architecture of PS2.0 is an extension of PS1.4. Clearly, the ATI's architecture has influenced DX9 PS. Many of the new PS2.0 instructions and registers have the same names as their PS1.4 counterparts, whereas many of the old tex* DX8 PS1.0 instructions have been deprecated.
Wrong! That's all I can say... :wink:
Dave Baumann
07-Jun-2002, 16:05
But I'd say that vertex shaders were done (almost?) completely by NVidia (OpenGL extensions behave exactly the same as VS.1.1). The only difference between VS.1.0 and VS.1.1 is that VS.1.1 has address register and VS.1.0 does not
Surely, though, Radeon 8500 was already too far in development for anything NVIDIA had one for NV20 to have had any influence?
I thought the PS 1.x was IHV (vendor) specific. My minds fading, but wasn't 1.3 nVidia, 1.0 or 1.1 3dfx?
MDolenc
07-Jun-2002, 16:31
As I said there is no hardware that supports only PS.1.0. GeForce 3 supports PS.1.0 and PS.1.1. GeForce 4, Parhelia, Xabre added support for PS.1.2 and PS.1.3. Radeon 8500 supports all PS.1.x shaders. Future hardware that will support PS.2.0 shaders will also support all PS.1.x shaders. 3Dfx never made any pixel shader capable part (maybe that's what they named "texture computer").
Doomtrooper
07-Jun-2002, 16:43
Ps 1.1 is Nvidia (Geforce 3-X-BOX), then Came Ps 1.4 for Dx 8.1 which was ATI(Radeon 8500-Highest PS Support @ 1.4), then came Nvidia again with PS 1.3 support with a Geforce 4 :P
MDolenc
07-Jun-2002, 16:49
Doomtrooper: As I said, X-Box does not actually have PS.1.1. X-Box has much more "OpenGL register combiner like" pixel shaders.
Doomtrooper
07-Jun-2002, 16:55
I didn't know that, I thought that is why Microsoft was pissed @ nvidia about having to buy license agreements for Pixel Shader rights in the X-box...
DemoCoder
07-Jun-2002, 18:52
Come on guys get serious...
PS.1.4 has one arithmetic instruction more then PS.1.3 and only 4 texture addressing instructions. Texture addressing and arithmetic instructions can be mixed in PS.1.4.
PS.2.0 is of course way ahead with 64 arithmetic instructions and 32 texture addressing instructions and again two phases...
PS1.4 introduced the idea of "phases" and decoupled iterators from samplers. This architecture remains in PS2.0. If PS2.0 was just PS1.1 or PS1.4 with more registers, addressing modes, and longer programs, it would still suck.
The uncoupling of the tex* instructions so that now you can separately perform math on the texture coordinates, and separately sample them into other registers, is an enormous improvement. The ability to have a phase in the pixel shader dedicated to doing texture coordinate math and dependant reads is the improvement to the basic architecture that PS1.4 added and that remains in 2.0. In PS < 1.4, you had no ability to mix math instructions in the tex* stages. Of course, NVidia's hardware lacks the ability to do this.
You can count instructions all you want and while it is true that PS1.4 only added the texld and texcrd instructions, it is the new architecture introduced that matters more.
No one in this thread is trying to argue who "invented" pixel shaders. We are talking about politics. Of all the hardware vendors, who has most influence over the DX9 pixel shader spec It is clear that DX8 was heavily influenced by NVidia. NVidia has patents on the shaders. Some instructions map 1-1 very specifically with NVidia's NSR and not with the old TSS. PS1.2/PS1.3 even more so, as Nvidia attempted to expose more of what you can do in the OpenGL extensions within PS.
Now all we are saying is that PS1.4 is clearly influenced by ATI's hardware, and if you look at PS2.0's overall *architecture* it is built off of PS1.4's, using phases and separate samplers, not the old PS1.0-1.3 style. Moreover, in the past, Microsoft has played this game of playing vendors against one another using each DirectX release. Why would they stop with NVidia?
By the way guys, is the GF3 PS1.3 capable or not?
I've heard varying accounts on that one, some have said that later drivers exposed 1.3 capabilities, others say only GF4 is 1.3-capable and that GF3 hence is 1.1-only (or maybe 1.2).
If anyone could clear that up it would be appreciated...
*G*
MDolenc
07-Jun-2002, 20:37
I just wanted to say that there are still all those PS.1.0-PS.1.3 tex* instructions in PS.2.0 that were not present in PS.1.4.
But I guess I should stress the differences that PS.1.4 made...
And for the politics: I think it is quite obvious that ATI influenced PS.2.0 most, but that does not mean that NV30 won't support them. From what I have heard NV30 is going to go over PS.2.0. And if this is correct then we might see DX 9.1...
And if this is correct then we might see DX 9.1...
I have heard this rumor about possible DX9.1 also and I think that another fragmentation of API-s, cards and versions are quite the contrary what the industry needs right now :cry:
But the business is business and MS has to push their "consistent" XBox platform somehow...
Joe DeFuria
07-Jun-2002, 21:38
I don't think it's what we need right now either...but it would be fun to watch the nVidia and ATI camps "flip-flop" on the DX support issue.
ATI: What do we need DX 9.1 for...it will never get support because too few cards support it. The installed base is DX9.0, and there's nothing you can do with 9.1 that you can't do with 9.0 anyway.
nVidia: DX 9.1 offers a large advantages over DX 9.0. Developers praise it, and if you don't have it, you'll be missing out.
It would also be interesting to see how MADONION would handle such a situation. Does anyone want to predict about when we'll see 3D Mark 2002? Close to the ATI launch...or the nVidia launch? :-?
Geeforcer
07-Jun-2002, 21:40
It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention.
On the contrary. The Radeon 8500 chip, AFAIK, is a great seller for ATI. I'm sure ATI's board partners want as the R-300 product launch to be as close as possible to actual shipping. Otherwise, they can be left holding lots of Radeon 8500 stock, and forced to make price cuts to unload them.
Actually, that would make perfect sense, considering that 8500 is almost a year old and no longer has the performance crow. By hyping R300, ATI could negatively impact the sales of the fastest card right now, GF4. Nvidia cannot do that since their latest card was released only a few months back and still has to pay for itself. Considering that 80% of the general public's buying decisions would not be effected anyway and the remaining 20% of enthusiast might be compelled to wait instead of upgrading to GF4 this summer hyping R300 would be a smart move for ATi. This time last year, they were actively hyping R200.
Is anyone want to predict about when we'll see 3D Mark 2002? Close to the ATI launch...or the nVidia launch?
I'd say that they try to keep their word and release their Mark soon after DX9. They can always release a patch for newer cards and somehow I think that patch for NC30 support will not be in works as long as Kyro's was :lol:
Joe DeFuria
07-Jun-2002, 21:50
By hyping R300, ATI could negatively impact the sales of the fastest card right now, GF4.
They would also negatively impact the sales of Radeon 8500. Radeon 8500 prices are already adjusted below GeForce4 prices. 8500 is not the fastest card, but it's not the most expensive either.
You are also forgetting my other point: ATI has OEMs to keep happy. OEMs aren't making as much profit on Radeon 8500 as they would if GeForce4 wasn't around, but they would make even LESS (or none) once R-300 is announced.
It's not as simplistic as "ATI doesn't have the fastest card, so they should hype their next card."
....and the remaining 20% of enthusiast might be compelled to wait instead of upgrading to GF4 this summer hyping R300 would be a smart move for ATi.
Again: ATI Hypes R-300 = nVidia OEMS lower price of GeForce4 due to less demand = ATI OEMS must lower price of Radeon 8500 because of GeForce4 price drop = ATI Oems not happy.
This time last year, they were actively hyping R200.
Certainly not. The first we heard of R-200 was in August. (We hear about the "Smartshader technology" in July.)
But again, LAST YEAR ATI was not selling parts to OEMs. So ATI only had "themselves" to consider when introducing their next part. R-200 and RV-200 were the first parts sold to 3rd party AIB manufacturers, so they would actually WELCOME early hype from ATI...they had no prior ATI parts that would lose sales.
RussSchultz
07-Jun-2002, 21:53
ATI: What do we need DX 9.1 for...it will never get support because too few cards support it. The installed base is DX9.0, and there's nothing you can do with 9.1 that you can't do with 9.0 anyway.
nVidia: DX 9.1 offers a large advantages over DX 9.0. Developers praise it, and if you don't have it, you'll be missing out.
Then followed shortly by:
ATI: You definately need DX9.2. It is so much better than DX9.1, you'll be missing out.
nVidia: DX9.1 is the standard, nobody out there is using DX9.2. Don't be on the bleeding edge!
(and of course)
SiS: We're cheap! Buy our card please!
Matrox: We're canadien! Buy our card please!
Videologic/ImgTec: We're cool! Build our card, please!
Bitboys: We're desperate! Invest in us, please!
Doomtrooper
07-Jun-2002, 22:02
By the way guys, is the GF3 PS1.3 capable or not?
I've heard varying accounts on that one, some have said that later drivers exposed 1.3 capabilities, others say only GF4 is 1.3-capable and that GF3 hence is 1.1-only (or maybe 1.2).
If anyone could clear that up it would be appreciated...
*G*
http://discuss.microsoft.com/SCRIPTS/WA-MSD.EXE?A2=ind0109C&L=DIRECTXDEV&P=R9271
Randell
07-Jun-2002, 22:15
ROFL @ Russ :D
multigl2
07-Jun-2002, 22:18
>>"There was actually so called PS.0.5 in DirectX 8 Beta that Radeon and GeForce 1/2 supports, but it was dropped before DirectX 8 made it public. NVidia had drivers that supported this version but dropped support when Microsoft decided that is does not want PS.0.5, but ATI renamed it into 1.0 but again dropped support latter."<<
Nope, no such thing as PS 0.5, ATi people have stated this many times.
Geeforcer
07-Jun-2002, 22:22
Certainly not. The first we heard of R-200 was in August. (We hear about the "Smartshader technology" in July.)
But again, LAST YEAR ATI was not selling parts to OEMs. So ATI only had "themselves" to consider when introducing their next part. R-200 and RV-200 were the first parts sold to 3rd party AIB manufacturers, so they would actually WELCOME early hype from ATI...they had no prior ATI parts that would lose sales.
Wrong on both counts: ATi started selling chips to OEMs as well as announced Trufrom back in May last year.
LeStoffer
07-Jun-2002, 22:54
From what I have heard NV30 is going to go over PS.2.0.
I only heard this rumor about DX 9.0 and DX 9.1 before on www.theinquirer.net or www.theregister.co.uk so could you please elaborate a bit about what you heard. I think that a lot of us would like to know.
umm... anyone remembers PS 1.0 / VS 1.0 ??
they were for DX8 and made by 3dfx, if I am not mistaken...
Yes, PS 1.0 was for 3dfx' Rampage, 1.1 is Nvidia's Geforce3, 1.3 is Geforce4 and 1.4 is Radeon 8500.
Sabastian
08-Jun-2002, 00:02
umm... anyone remembers PS 1.0 / VS 1.0 ??
they were for DX8 and made by 3dfx, if I am not mistaken...
Yes, PS 1.0 was for 3dfx' Rampage, 1.1 is Nvidia's Geforce3, 1.3 is Geforce4 and 1.4 is Radeon 8500.
Err, I don't think that your statement about the gf3+4 is correct. I was under the impression that the gf4 had the same ps capasity as gf3. I may be wrong and will take being corrected happily.
Sabastian
Chalnoth
08-Jun-2002, 00:46
Nope, the GeForce4 has some additional shader ops, mostly revolving around dependent texture reads. This is why none of the GeForce4 demos operate properly on a GeForce3.
Ailuros
08-Jun-2002, 03:23
umm... anyone remembers PS 1.0 / VS 1.0 ??
they were for DX8 and made by 3dfx, if I am not mistaken...
Yes, PS 1.0 was for 3dfx' Rampage, 1.1 is Nvidia's Geforce3, 1.3 is Geforce4 and 1.4 is Radeon 8500.
Although not 100% certain I recall for Spectre:
Vertex Shader 1.0 (lacking address ops)
Pixel Shader 1.1
Then followed shortly by:
ATI: You definately need DX9.2. It is so much better than DX9.1, you'll be missing out.
nVidia: DX9.1 is the standard, nobody out there is using DX9.2. Don't be on the bleeding edge!
(and of course)
SiS: We're cheap! Buy our card please!
Matrox: We're canadien! Buy our card please!
Videologic/ImgTec: We're cool! Build our card, please!
Bitboys: We're desperate! Invest in us, please!
ROFL Russ :D
I'm a fan of sarcasm not w/o a reason: most effective way to illustrate reality.
Joe DeFuria
08-Jun-2002, 04:58
Wrong on both counts: ATi started selling chips to OEMs as well as announced Trufrom back in May last year.
First of all, ati made a PR about TRUFORM in May, but it was not to directly hype Radeon 8500. Truform was supported in DX8, and ATI enabled it in SOFTWARE in their Radeon drivers.
Second, as I said, ATI did not launch the chip until August. Did we know it was coming? Sure. Just like we NOW know the R-300 "is coming" based on ATI press releases about "next-gen parts" being demoed on various software and workstations.
The point is, at "this time last year", ATI did NOT HYPE the Radeon 8500 in terms of product specs or timing.
ATI is in fact "hyping" R-300 to a similar degree at this time, compared to Radeon 8500. And that is...not much. Nor should they do any hype until they are much closer to actual shipping...like one month away.
And ATI officially launched their OEM strategy in June, and I still believe that the R-200 and RV-200 chips were among the FIRST that OEMs actually bought. Certainly, OEMs were not involved in buying original Radeon chips during that chip's original launch.
Chalnoth
08-Jun-2002, 05:41
First of all, ati made a PR about TRUFORM in May, but it was not to directly hype Radeon 8500. Truform was supported in DX8, and ATI enabled it in SOFTWARE in their Radeon drivers.
Heh, you don't call that hype? They publicly stated that their next card would do Truform with no speed hit...
Doomtrooper
08-Jun-2002, 05:50
There IS little speed hit when the engine isn't truforming everything on the screen, RTCW with Truform enabled..maybe 2 fps..Serious Sam without exclusions and very fat trees, birds, fish and whatever else is on the screen, terrain, weapons...HUGE HIT..as expected :-?
Geeforcer
08-Jun-2002, 09:08
First of all, ati made a PR about TRUFORM in May, but it was not to directly hype Radeon 8500. Truform was supported in DX8, and ATI enabled it in SOFTWARE in their Radeon drivers.
Second, as I said, ATI did not launch the chip until August. Did we know it was coming? Sure. Just like we NOW know the R-300 "is coming" based on ATI press releases about "next-gen parts" being demoed on various software and workstations.
The point is, at "this time last year", ATI did NOT HYPE the Radeon 8500 in terms of product specs or timing.
ATI is in fact "hyping" R-300 to a similar degree at this time, compared to Radeon 8500. And that is...not much. Nor should they do any hype until they are much closer to actual shipping...like one month away.
And ATI officially launched their OEM strategy in June, and I still believe that the R-200 and RV-200 chips were among the FIRST that OEMs actually bought. Certainly, OEMs were not involved in buying original Radeon chips during that chip's original launch.
Joe, if Truform announcement does not fall under hyping, then I don't know what does. Demos, press briefings, headlines like: "ATI TRUFORM Technology - Powering the next generation Radeon" - seems a lot like hyping to me. They are defiantly hyping R300 right now; just look at Doom3 and Intel "next generation" press releases. That's a lot of PR covering an unannounced card, in fact I challenge you to find his kind of official PR regarding unannounced next gen chip from any other graphics company. I don't see how anyone could look at all of this and say that ATi didn't/doesn't hype their producst, especially considering that there is nothing wrong with them doing that - actually it is a good strategy.
Chalnoth
08-Jun-2002, 09:20
Actually, I think the R300 hype is almost nonexistant...the DOOM3 thing was just a bonus for ATI...
Doomtrooper
08-Jun-2002, 16:56
Actually, I think the R300 hype is almost nonexistant...the DOOM3 thing was just a bonus for ATI...
Hmmm ATI sponsoring QuakeCon 2002 and offering $100,000 in prize money seems like a pretty good way to HYPE :wink:
http://www.quakecon.org/
Where is Typedefenum, he was saying to ATI-ISV right here on this forum ATI needs to get involved more with the gaming community..well here is another step ATI has done.
Geeforcer
08-Jun-2002, 17:13
Actually, I think the R300 hype is almost nonexistant...the DOOM3 thing was just a bonus for ATI...
I disagree. Again, they have staged several high-profile demonstrations and have several press releases covering their unannounced product. (here (http://www.ati.com/companyinfo/press/2002/4499.html) and here (http://www.ati.com/companyinfo/press/2002/4495.html)). They are taking the spotlight off GF4 and putting it back on ATi, a great strategy IMO.
I dont see that kind of stuff making much impact on end users ... something like the P10 technology description can create some hype, but some 0 information press releases are just ignored as they should be. Ill have to agree the DoomIII was a coup, mostly because it constituted at least some kind of endorsement, but for the rest nothing much is registering.
Geeforcer
08-Jun-2002, 17:21
One look around the Internet message boards clearly shows that they are not ignored.
Chalnoth
08-Jun-2002, 18:06
People only thought it was spectacular that DOOM3 was running on an ATI video card when they thought that nVidia was trying to get id software to run an NV30. Now it's just, "Oh, that's kind of cool, but I already knew it should be able to do that!"
Of course, there is the speculation that the R300 will be out first, but I don't see how that helps ATI, since nobody will really care until the cards are actually released.
Besides, we do know that it will be at least August before these cards are to be released, and probably September or later before they're available, and as a result, I don't really think I would recommend people waiting just yet, particularly if they want to purchase on the mid-low range. These high-end cards won't have much impact on the <$200 market.
Typedef Enum
08-Jun-2002, 18:31
One thing that does give you a good hint about R300's release is the following:
The QuakeCon attendees will be among the very first to get their hands on our next-generation product - we're confident they will be blown away
That event takes place August 15-17. Should that hold water, it would appear that reviews will come out in/around the same time, and R300 should make its way into retail around September....October @ the very latest.
With that said, it seems virtually certain that NV30 will end up following the R300 release, unless they end up doing a Star-Wars kinda deal...("double your efforts.", with regards to building the Death-Star :) ).
Chalnoth
08-Jun-2002, 18:57
Of course, that would also tie in closely with nVidia's releases of the past...
Typedef Enum
08-Jun-2002, 20:01
True...but it certainly doesn't seem very likely this time around. Unless nVidia wants to just completely lay some smack on people out of the blue, we haven't heard so much as a hint about NV30 being shown @ any function...be it behind closed doors, Asia, developers, etc.
On the other side of the coin, R300 has made its way into many corners. I think NV30 will likely end up a October/November release...But they will probably unveil the chip specs in August.
Joe DeFuria
09-Jun-2002, 01:41
They are defiantly hyping R300 right now; just look at Doom3 and Intel "next generation" press releases.
GeeForcer,
As everyone else is trying to tell you, these types of PR don't tell us much more than we already know: The R-300 is coming.
Do we need a PR from Intel to tell us that the next Pentium chip is on its way? Is that the kind of thing that drives current Athlon prices down?
What constitutes the kind of hype that prompts price drops is the following:
* Reveal clock speeds / performance
* Reveal expected availability
* Reveal expected price.
That's what typically comes with an official product "launch". And even THAT won't drive prices down unless price/performance is competitive, and availability isn't "too" far away.
That's a lot of PR covering an unannounced card, in fact I challenge you to find his kind of official PR regarding unannounced next gen chip from any other graphics company.
Lol...
http://www.nvidia.com/view.asp?IO=IO_20020109_4391
nVidia announcing hardwired T&L about 6 months before GeForce256 is launched. :roll:
And I won't even get into 3dfx's "Napalm" technology launch which was also months before the VSA-100 boards were announced....
Ailuros
09-Jun-2002, 03:00
I didn't see a technology announcement from ATI on their next part like Matrox f.e. presented Displacement Mapping a while ago.
Napalm suffered some major delays by the way.
I'm not aware of any other company that "indicates" that developer X for example is running on next generation hardware before the said product has been even announced.
But in the end too much fuzz over nothing. I couldn't care less what each company touts until I see a product tested/analyzed.
Geeforcer
09-Jun-2002, 04:16
GeeForcer,
As everyone else is trying to tell you, these types of PR don't tell us much more than we already know: The R-300 is coming.
First off, "everyone else"? Huum, that would be Chalnoth. No need to get a non-existent group consensus involved into this. Second, these PRs do tell us some other things, or at least ATi's interpretation of them:
ATI's chip was chosen because it offers the most advanced, high-performance graphics architecture to power the demonstration at E3 2002. DOOM III™, 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 III™ 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 III™ with ATI's next generation hardware - they'll live it."
What they are saying is not merely "R300 is coming" but "R300 will be a great card for DoomIII", which so happens to be the most highly anticipated PC game in recent history. The whole DoomIII thing was a huge shot-term PR victory for ATi and their upcoming product.
What constitutes the kind of hype that prompts price drops is the following:
* Reveal clock speeds / performance
* Reveal expected availability
* Reveal expected price.
I never said anything about “hype that prompts price drop”. I was talking about “hyping” in general and the definition of "hyping" is "to publicize or promote". Are you going to deny that DoomIII and Intel demos as well as the corresponding press releases were meant to do just that, publicize and promote upcoming card? Because that would constitute hyping.
Lol...
http://www.nvidia.com/view.asp?IO=IO_20020109_4391
nVidia announcing hardwired T&L about 6 months before GeForce256 is launched. :roll:
This is along the lines of TruForm/Smartshader/FSAA PR is not what I specifically asked for. What I asked for were public demonstrations of unannounced next-generation hardware, analogs to DoomIII demo. So my challenge still stands: "...find this kind of official PR regarding unannounced next-gen chip from any other graphics company."
And I won't even get into 3dfx's "Napalm" technology launch which was also months before the VSA-100 boards were announced....
So 3dfx publicly demoed a high-profile game on Napalm prior to it's announcement and then issues a press release to cover it? I must have missed that.
Back to the topic of "NV30 having roughly a whole crap load of transistors":
Is it possible that Nvidia is trying a 3dfx move with the T&L chip off board, and when Nvidia's CEO says there are 120m transistors, he means between the 2 chips?
Probably not...but it seems semi-feasible.
Chalnoth
09-Jun-2002, 05:04
The 2nd T&L unit is not happening.
Geeforcer
09-Jun-2002, 05:05
Is it possible that Nvidia is trying a 3dfx move with the T&L chip off board, and when Nvidia's CEO says there are 120m transistors, he means between the 2 chips?
Unlikely: if that was the case, .13um would not be such a big issue for Nvidia.
Has anyone confirmed that the latest wired magazine has the interview?
I have the Wired´s june edition with Steven Spielberg in the cover. Nothing about nvidia.
Doomtrooper
09-Jun-2002, 20:34
Searching the issue contents shows nothing, even going back a couple of months...
http://www.wired.com/wired/current.html
Has anyone confirmed that the latest wired magazine has the interview?
I have the Wired´s june edition with Steven Spielberg in the cover. Nothing about nvidia.
It's the current issue, red with Huang on the cover.
So have you seen it? Do you mean the july issue?
Hellbinder
10-Jun-2002, 03:12
Chalnoth,
One question. Do you even live on planet earth? I have never seen such an unadulterated continuous spewage of unballanced, untrue, misguided, and down right totally one sided series of posts. You completely ignore the obvious truth, or even common sense.
To think you are a part of at least a "somewhat" reputable website.
unbelievable :roll:
Joe DeFuria
10-Jun-2002, 03:28
GeeForcer,
I don't understand where you're coming from. This whole debate stems from OopmaLumpa's oringinal comment:
It is strange that ATI hasn't been more forthcoming. It's easy to understand NVIDIA's reluctance to cannibalize sales, but ATI really should be trying to get some attention.
So let me try and summarize the aruments so far:
Oppma: He can't understand why ATI isn't "hyping" R300, for the purpose of impacting sales of nVidia chips. Then I said ATI shouldn't hype it. (Both of us seem to agree there is no real "hyping" going on in the sense that it could impact sales of GeForce4).
Then you jumped in an said ATI "should" hype it....
And now you're saying they "are" hyping it?
So, erhm...which is it? Are you saying that ATI "is" hyping their part to the extent that they are trying to actively lower the sales of todays high-end chips like GeForce4 and Radeon 8500, or are they not hyping it, but should?
First off, "everyone else"? Huum, that would be Chalnoth.
No, every person (sure, it's only a few ;)) who has cared to comment on this debate one way or the other (whether they think it "should" be hyped or not), doesn't consider the R-300 PR to this point to have any major impact. That would be:
OompaLoopa
mfa
Chalnoth
I have not seen anyone agree that ATI's PR to date is designed to get people to think about not buying a Geforce4. They may be out there, but they haven't said anything.
What I asked for were public demonstrations of unannounced next-generation hardware, analogs to DoomIII demo.
I have no idea what your point is. This is not the first time that un-announced hardware was "confirmed" officially by a company as my past examples indicated. How does "confirming" the existence of the next part by running a Doom3 demo mean that this is some level of hype above and beyond other "non product launch" hype?
So 3dfx publicly demoed a high-profile game on Napalm prior to it's announcement and then issues a press release to cover it? I must have missed that.
I don't recall if they used some actual games, but they did have the t-buffer demos, with a complete press EVENT to cover it, long before the products were actually announced.
So have you seen it? Do you mean the july issue?
Yes, I read the article, and the 120M figure was about the extent of the "new" information concerning future graphics cards.
Geeforcer
10-Jun-2002, 04:24
Besides you, Chalnoth was the only person to address anything that I've said in this thread; however this is not very relevant.
My points were:
1) ATI should hype R300, because IMO they would hurt Nvidia much more then it would hurt them. Since Nvidia has the speed crown ATM and GF4 is fairly new and thus has not sold enough units to be overly profitable, Nvidia would be hurt the most should consumers who where planning to upgrade this summer would decide to postpone their purchase till fall in light of impending R300 arrival. (IMO, they are already doing this, see further below)
2) In the past, ATi has actively hyped (promoted) their cards well in advance by showcasing the underlying technology. Truform press event was held in May, followed by Smartshader unveiling several weeks later; both were intended to promote R200 - fairly successfully I might add, as they managed to rise the interest in ATI's fall product as well as generate quite a few discussion around the net.
3) The change in business model would not necessarily discourage ATi from starting technological unveiling/promotion campaign well in advance. Last summer, ATi was already selling chips (R100) to OEMs while in the same time promoting the yet-unannounced R200. It doesn't mean that they would do the same thing last year (ATi relationship with OEMs has greatly expended in the past year) but in the past OEM commitments did not preclude ATi from promoting unannounced chips in the past.
4) To a degree, ATi is already seeking to generate press coverage and public interest in their upcoming unannounced product by staging high-profile demonstrations. DoomIII demo in particular changed the way many people view ATi; not only were they involved in perhaps the most significant demo of E3 covering one of the most highly anticipated PC games in recent history but they also their upcoming card was praised by Carmack at the expense of their arch-rival. That is especially relevant considering that he had some not-so-nice things to say about their current product, in a way it can be viewed as a redemption and certainly had a significant impact on the hearts&minds of some people.
Does this hype result in the sales decline of current products? That would be very hard to say without looking at very recent market data, which I don't have. However, I am willing to bet that there are quite a few people right now thinking: "Should I go ahead with my plans and get that GF4 4600, considering that early version next ATi's card is faster in DoomIII and is just a few months away?"
Johnny Rotten
10-Jun-2002, 05:37
]Chalnoth,
One question. Do you even live on planet earth? I have never seen such an unadulterated continuous spewage of unballanced, untrue, misguided, and down right totally one sided series of posts. You completely ignore the obvious truth, or even common sense.
To think you are a part of at least a "somewhat" reputable website.
unbelievable :roll:
"Hello kettle, this is pot calling. You're black." :roll:
Typedef Enum
10-Jun-2002, 05:58
Ain't that the truth baby...
Althornin
10-Jun-2002, 17:23
Ain't that the truth baby...
Hypocracy doesnt make his point invalid.
So have you seen it? Do you mean the july issue?
Yes, I read the article, and the 120M figure was about the extent of the "new" information concerning future graphics cards.
Thanks Brem. I will wait the magazine.
1) ATI should hype R300, because IMO they .....
Hmmm yea but they really have to be carefull as too much hype or hype that fails to deleiver is very bad press. So thats the challange of the PR department. Come up with just the right about of hype. Too much = Smoothvison as that kind of failed to live up to its perfromance where as the Turform/smartshaders are a good example of good hype.
dominikbehr
10-Jun-2002, 22:14
Both ATI and nVidia are not in business of "hurting" each other. They are in business of making and selling graphic chips (and increasing shareholder value). Please tell me how "hurting" yourself and "hurting" other company little more increases shareholder value?
Geeforcer
10-Jun-2002, 22:44
Both ATI and nVidia are not in business of "hurting" each other. They are in business of making and selling graphic chips (and increasing shareholder value). Please tell me how "hurting" yourself and "hurting" other company little more increases shareholder value?
Obviously they are not going to try and "hurt" each other just for the hell of it, but to increase the market share it the expense of the rival. By the same token, how does occasional low-brow PR aimed at each other increase shareholder value?
Chalnoth
10-Jun-2002, 22:58
Sometimes self-control is just very, very hard. When you have something you think is great (or think will be great), it's all too easy to show it off before it's entirely ready.
dominikbehr
10-Jun-2002, 23:07
Geeforcer:
Obviously they are not going to try and "hurt" each other just for the hell of it, but to increase the market share it the expense of the rival. By the same token, how does occasional low-brow PR aimed at each other increase shareholder value?
Again, both ATI and nVidia are not working to get bigger market share, they are working to sell more chips/cards. Any PR hurting yourself, your rival and whole market just to gain imaginary markeshare is very risky at least. Usually PR and marketing tries to sell more stuff.
Btw, cable and DSL companies do not seem to understand this.
Hellbinder once said:
give me a break guys....
he is OBVIOUSLY talking about the Nv30. The R300 is not due to be released for severl moths. It doesn't even make LOGICAL SENSE to assume he is comparing the next generation to this generation. The next generations superiority is an OBVIOUS GIVEN.
No he did not say it specifically. But the intent is clear, and the reason he said it is clear. He HAD to put to rest all the BS rumors that Nvidiots started spreading.
in thread http://www.beyond3d.com/forum/viewtopic.php?t=1101
And yet, here he is posting:
One question. Do you even live on planet earth? I have never seen such an unadulterated continuous spewage of unballanced, untrue, misguided, and down right totally one sided series of posts. You completely ignore the obvious truth, or even common sense.
To think you are a part of at least a "somewhat" reputable website.
unbelievable
as if his opinion is worth anything.
I guess it's his "are you totally fukcing STUPID?!?" attitude that I dislike.
Chalnoth
11-Jun-2002, 00:01
Again, both ATI and nVidia are not working to get bigger market share, they are working to sell more chips/cards. Any PR hurting yourself, your rival and whole market just to gain imaginary markeshare is very risky at least. Usually PR and marketing tries to sell more stuff.
Btw, cable and DSL companies do not seem to understand this.
Don't forget that there's also the fact that people need to know about your product in order to buy it. If you put out tons of PR, and can't deliver, is it really worse than just failing to deliver? In some cases, I'm sure it is, but not in all.
dominikbehr
11-Jun-2002, 00:36
indeed, bullsh*t PR usually fires back.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.