NV25 details, release date Spring 2002

[noko]

Looks like Nvidia is backing off the 6 month release period for major chip upgrades which brought them to the #1 graphic chip supplier position. A second vertex shader like the XBox GPU and a basic speed up with potential DX 8.1 ability. Also looks like the Radeon 2 has a good chance of reigning surpreme for awhile with its expected 250mhz GPU with a more advance Pixel Shaders ver 1.4 vice the 1.1 version of the GF3. Also ATI is able to make a Radeon 2 Max combo with there newly designed bridge chip to coordinate the drawing of frames. Still shrouded in mystery is the specs for the Radeon 2 which makes allot of us wonder why they don't start bragging a little and start raking in some money with pre-orders. There are always plenty of early birds with money falling out of their pockets to grab a few bucks from at elevated prices approaching space travel.


<A HREF="http://www.theinquirer.net/06070118.htm" target="_new">http://www.theinquirer.net/06070118.htm</A>
<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/06/01 03:06 PM.</EM></FONT></P>

 

[rcf84]

Yup, I'm waiting for the Radeon 2 Maxx the NV25 Killa.

Or should i wait for the Radeon 3 the nvidia finisher.

Nice Intel and AMD users get a Cookie.... Yummy

 

[noko]

Makes me think that the NV30 will be a 2003 release chip vice a late 2002 release. So is Nvidia really slowing down what they started, could be devastating with STMicro ready to fill where 3dfx left off. Then again the NV30 could be a totally new chip design from the ground up which I hope it is. So a long delay with a major design change may be worth it in the long run. Plus how well will Nvidia integrated chipset (NForce) sell? The general slow down will affect everyone and the prices for all chipsets will drop like rocks from the sky meaning Nvidia may have to cut profits to get the nforce selling well. Now the question is will ATI have an Intel integrated chipset available soon? Still Intel sells more then AMD chipsets so in this case ATI could be in a more favorable position if they deliver a viable integrated chipset. The NForce may just nudge Intel to get ATI to speed up their new chipset, hopefully it will be more powerful then the feeble mx200 capable NForce.

 

[rcf84]

Yeah ATi has the i810 for the P4. There chipset support DDR-SDRAM w/ UMA + Intergrated Rage128 pro upto 32mb !!! Also the R300 is built ground up too !!! When the NV25 is out the Radeon 3 will be out with fully Directx 9 ready !!

Nice Intel and AMD users get a Cookie.... Yummy

 

[noko]

Rage128pro? thats not so hot in itself since the Nforce is a T&L capable graphics engine. DDR would be a very nice change, we will see when it is release how good or bad it is. Still sounds like the Nforce will be a better chipset then the ATI chipset if that is the case but then the ATI chipset will probably sell more due to it being Intel capable. Not even mentioning the sound capability of the Nforce.

 

[rcf84]

The thing is its Intel's i810 for the P4. Intel will shove into all the OEM markets. ATi chipset could use the ICH2 southbridge rumor???

Nice Intel and AMD users get a Cookie.... Yummy

 

[HolyGrenade]

Those bastards. I don't understand it. If they were going for a late release, why not just skip this junk altogether and distribute more resources to the nv30. Or is resourcing for the nv30 the real reason for a late nv25 release?

So is this nv25 the GF3 ultra or not? I suppose they could just release a clocked up nv20 with faster ram and call it GF3 Pro.

Do you think ATI will take nvidias decision into account and follow suit, by enjoying their current sales and conserving resources? or, do you think they'll take this opportunity to jump the gun and go for an early release?

I wonder what the kyro boys are up to. What do you think they make of this change of staratagy nvidia just made.

If you think about it, it does seem to make a bit of sense. All these years they enjoyed huge vertical growth. Since they entered the game, they've never ceased to impress. With their TNT series, they started to lead. But now they realise, they'll soon hit the ceiling. The different between a desktop graphics card and a super duper workstation graphics card is far lower than it used to be. So, they are starting their horizontal growth, which inevitably means a slower vertical growth. Their spreading into all sorts of other areas.

During the vertical growth, they've seen of several ancients as well as fellow newbies. ATI and matrox seem to be the only big players to have survived. They,ve survived through their financial might and their market stronghold (or maybe it was just a foothold ), nVidias fellow newbie, Videologic, had to start early horizontal growth to survive.

Now, nVidia and ATI have started their horizontal growth, which seems to leave matrox struggling. Perhaps they'll try to claim the digital home video market. Don't worry I haven't forgotten the others. They are hanging around in the value market or in integrated graphics on motherboards.

This leaves me to worry about creative labs. I kinda like them. I know they have a multimedia solution on the horizon. But beyond that?


P.S. Apologies if you feel I've drifted off the subject.


<font color=red><i>Tomorrow I will live, the fool does say
today itself's too late; the wise lived yesterday<P ID="edit"><FONT SIZE=-1><EM>Edited by holygrenade on 07/07/01 00:24 AM.</EM></FONT></P>

 

[rcf84]

Well ATi just got FireGL and there Intel P4 chipset will get alot of sales with intel backing. Well ATi has been here for a long time with matrox and Trident. ATi is the last of the Older companys left still at tops. remember Nvidia and 3dfx when they first start guess what ATI and thats right s3 where battling for No.1

Nice Intel and AMD users get a Cookie.... Yummy

 

[HolyGrenade]

Which part of that haven't I already said?


<font color=red><i>Tomorrow I will live, the fool does say
today itself's too late; the wise lived yesterday

 

[noko]

Hopefully what Nvidia is doing is designing a completely new chip with the NV30. Something revolutionary vice evolutionary. The GF3 actually competed against Nvidia GF2 Ultra while ATI was making money by selling their cards below $200. Getting their sells back from the dismal Rage128 Pro sells. So Nvidia basically cut their own throat in a small way.

The real money is in OEM and lower end market, except now the MX is competing against the cheaper Radeons but also the mighty Kyro2. Kyro2 is more unexpected then anything taking away Nvidia sells from numerous board makers. Top that off ATI is also in the game which Nvidia virtually had a monoply on (graphic card makers). So the horizontal market that Nvidia built up in the last 2 years is being dismantled rather suddenly.

Nvidia chipset Nforce should help them out but remember that most systems are still Intel, so if ATI can tap into those sells then ATI could potentially make more money then Nvidia even with an inferior chipset.

As for the fruitfullness of the Xbox Nvidia made out initially of some 200million dollars from Microsoft, not sure of any continued success there. While Nintendo GameCube could be a real treasure chest for ATI.

The professional market Nvidia built up a signficant lead in the low end market share, ATI now has the better more performing card, we will see how that will benefit them in the future. Still I don't think much money is to be had in the professional market. Except in the professional market you can become more of a leader in what is or is not on a video card, more influence in API development.

ATI seems to be timing their products about right, at the local Best Buy one Visiontech GF3 has been sitting there for now over a month while the rack of Radeon64's where sold out, interesting. A $400 card doesn't sell to well against a $200 card there. The Radeon box is better with obvious more features which the consumers understand much better and a cheaper price. Now they just got in a bunch of GF2 Pros but are selling over $200 with no VIVO, needless to say the Radeon 64 are out selling them too. For awhile it was around 2/3 of the shelf space was dedicated to ATI, now about 1/2 is. They have MX400s costing more then the Kyro2 from Hercules as well. I just don't see Nvidia doing to well in the Retail outlets now. Before I use to see Leadtech, Creative and others producing Nvidia cards, not no more in the Retail stores.

Nvidia needs to really improve there low end cards where the real money is vice the top end cards that is mostly a status symbol. A GF3 MX that is better then the Kyro2 and Radeons is needed by them. So maybe we will see Nvidia win back some of the card manufacturers with newer and better low end chips. Limiting the MX to 64bit DDR and 128bit SDram was a bad decision. Should have been able to do 128bit DDR allowing that chip to last awhile. Now it is one of the slower chips to be bought.

 

[noko]

Oops, I didn't answer your question:
<font color=blue>

Do you think ATI will take nvidias decision into account and follow suit, by enjoying their current sales and conserving resources? or, do you think they'll take this opportunity to jump the gun and go for an early release?</font color=blue>

I think ATI is going to follow through with there schedule release date, sell as many Radeons as they can while the board makers start making Radeons. ATI can then concentrate more on the Radeon 2 while you may see some interesting Radeon combos coming out.

For example, the new Radeon cores are SE cores which overclock up to 250mhz. So don't be to surprise to see Radeons with 3.8ns SDram like the Leadtech MX400. Also you may see some super Radeons out there with 4ns DDR ram as well running at 250mhz/250mhz plus. So ATI would have a hard time competing against the inventive third party board makers so having the Radeon2 out will make them not competing against them initially. The Radeon core is a better core then the MX by a long shot, more features then the GF2 cores. A 250/250 Radeon would be one high performing board, I don't see why it wouldn't be produced. Now this is speculation but it does make sense.

Actually I think ATI and Nvidia is wary of STMicro more then each other. STMicro has the cost advantage meaning it is cheaper to use a Kyro chip then something else except maybe Nforce. If STMicro comes through with the Kyro3 then everything could be changed. Maybe NV30 is going to be a tiler after all.

I am very interested to know what improvements ATI did to HyperZ. If ATI can get the efficiency better then the need to covert to a tiler would be less, supposenly the Radeon2 will have a texture buffer, meaning the most used textures will stay in the buffer conserving bandwidth of the onboard memory. Kinda like a texture cache, which may mean if you overclock the GPU a speed increase would be bought.

Time will tell, the real dilemma is that the hardware has advance so much more then the software that it is not necessary to upgrade. Good games and game engines are taking 2-4 years to develope, meaning the graphic cards abilities are not even remotely being taped. Even T&L is still lightly used in games and only now maybe optimized for it. ATI will be promoting more features with the Radeon 2:
. . Hydravision - come in handy with dual displays and video out display cards
. . Mpeg4 - Onboard Mpeg4 decoding, which means it will use the quality enhance features that ATI uses in hardware decoding like 9bit error term vice the usual 8bit error term used by software. Much less cpu usage. Also adaptive de-interlacing would be used improving quality.
. . Truform - N-Patches that can enhance old and new games easily. Don't necessarily have to wait for new games to take advantage of this type of tessilation.
. . Pixel Shader version 1.4 - Maybe Carmack won't complain about this implementation of Pixel Shaders.

<font color=purple><i>Shader Tools and Case Studies
Jason L. Mitchell - ATI Research

In this session, we will show a series of pixel and vertex shader case studies, progressing rapidly from simple animation and lighting models to more advanced effects. This will begin with a one minute introduction to vertex shaders, followed by a rapid progression to skinned characters, procedural deformation, tangent space transformations and several others. All of these code samples will appear in an assembly language which supports familiar features like #defines, include files, variable declaration and macros (inlined subroutines). We will illustrate the power of this model in managing the explosion of shader permutations that can be created in a production environment. The shader assembler tools and a library of useful macros will be made available to the developers.
At the pixel level, we will focus on the advancements made to the pixel shading model in DirectX 8.1, specifically the 1.4 pixel shading model. We will use a series of case studies and samples to illustrate this progression. Key advancements illustrated will include the power of the RISC instruction set and flexible dependent texture addressing. We will start with simple per-pixel lighting from one or more lights. From there, we will extend to per-pixel specular exponents and Fresnel terms. We can then show off 3D volume visualizations with dynamic transfer functions, super position of dudv maps, per-pixel anisotropic lighting and a variety of other effects. We will also provide developers with the ShadeLab tool which allows them to explore pixel shaders visually without getting bogged down writing application and vertex shader code to feed the pixel shaders.
Intended audience:
The intended audience for this session is a 3D graphics programmer who has at least a cursory level of either vertex or pixel shader experience. Those who have merely tinkered with some SDK apps will get a sense of where pixel shaders are headed in the 1.4 model. Those who are writing pixel shaders in a production environment will gain insights into how the new pixel shader model applies to their applications and how the vertex shader programming tools can help them manage complexity and code reuse.
<A HREF="http://www.microsoft.com/mscorp/corpevents/meltdown2001/session.asp#studies" target="_new">http://www.microsoft.com/mscorp/corpevents/meltdown2001/session.asp#studies</A></font color=purple></i>

ATI financial report is also looking much better then it did last quarter. So the Radeon 2 should hit in September. Lets hope.

Man I can sure ramble on, sorry.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/06/01 10:41 PM.</EM></FONT></P>