it's always been about NV35, not GFFX

[megadrive0088]

I realize what i'm about to say is nothing new, and probably very obvious but it needs restating for those expecting so much with GFFX.

---like TNT and GF3 before it, the first chip of Nvidia's next architecture, NV30/GFFX, has had its difficulties, shortcoming and disappointments...while its an amazing new technology, the first implimentation should be only for the hardest of the hardcore Nvidia users. with word that the Ultra 500/500 version will never reach retail,
its time to start looking toward the release of what will ultimately become the "proper implimentation" of NV30, that is, the NV35.
---The NV35 should be everything that was ment to be in NV30, everything that the engineers wanted but couldn't get in. sure, the clock speed of NV30 is alot higher than it was ever originally going to be. the R300 put the pressure on to get the performance up. so for that reason, we shouldn't expect a huge clock increase with NV35., but a host of new features, more of those vertex math processing units (the NV3X is a highly modular design like P10 from what I've read), perhaps more instructions in the pixel shader.
as well as a new anti-aliasing unit (like NV25 had over NV20) -
---I have little doubt that Nvidia will be back on its feet with NV35, and sooner rather than later. look how fast Nvidia came roaring back with TNT2 after TNT...those where the glory days for Nvidia. NVidia could have had GF4Ti in the fall of 2001, 6 months after GF3 and look what an improvement that was. GF4Ti has been Nvidia's finest chip IMO. with R350 coming to a PC near you soon, and R400 on the horizon for fall, NV35 should launch anywhere from April to July with cards being out anywhere from May to August. since NV4X is early 2004 at best (probably fall '04)
Nvidia needs a stellar chip for the late spring or summer.

 

[speng]

I'd say, "It's all about the NV35 when Nvidia noticed ATI's R300 was a stellar competitor and would require a better chip from Nvidia"

The NV30 (GFFX) was supposed to be convincingly better than R300.

Speng.

 

[overclocked]

Yes the NV35 vill most certain be released a couple of month´s earlier but as i understand it the R350 should be Officialy launched the 12´mars and likely to be in stores just a month or so from when FX@400/400 has hit.
Where does this leave NV31 and NV34 as ATI covered the mid-range with DX9 and i assume RV350 vill replace all DX8 budget and low end cards with DX9, like the Radeon9000 series etc.
Prices has get down on th 9700pro cards and a regular 9700 is the best Price/performance part for now.
So too me atleast it seems like ATI if they can pull of their aggresive "roadmap" and have a matured 13nm+low-k design on the R400 i can´t really see Nvidia catching up on ATI under 2003.
Only time will tell.

 

[THe_KELRaTH]

Just because a company does something right one year doesn't mean they're going to do it over and over again. You can see that in any industry. It usually take just one person to start off that magic product and that person may not even be on the team now.

 

[overclocked]

THe_KELRaTH
Just because a company does something right one year doesn't mean they're going to do it over and over again. You can see that in any industry. It usually take just one person to start off that magic product and that person may not even be on the team now.

Do you mean that ATI did right and Nvidia wrong, and that will turn around
Or the opposite?
Of course maybe Nvidia takes ATI by BIG suprise but the way it has turned now and in the past it seem´s really that ATI can hold it during 2003.

 

[megadrive0088]

I see ATI becoming more & more Nvidia like (what Nvidia used to be like)
throughout 2003 and 2004, with the R300, R400 and R500 lines.
Nvidia will be struggling, for at least 2003, but never defeated like 3Dfx. they'll probably be back on top in 2005 when they decide to release a true deferred rendering architechure using their held GigaPixel technologies. Unless ATI severely faulters with R400 (which is not primarily designed by ARTX staff as its mainly the R200 team behind R400) well thats just wild speculation. for all i know, the NV35 is a monster like GF2GTS or GF4Ti.

 

[g__day]

I disagree, I believe its about the third iteration that they get it honed, so by thinking NV35 will be a tweak that will find itself hard against teh R350 and it'll be NV40 vs R400 that'll be the showdown of real heavywieghts.

 

[Dave H]

It's very doubtful that the NV35 will be a significant departure from NV30 in terms of adding new execution units or new features. OTOH if there are significant features present but not properly functioning in NV30 silicon (which might be a good guess) they will probably be fixed and working in NV35.

 

[Nagorak]

That's all well and good... Unfortunately the R300 rocked. With ATi it wasn't "about the R350" (maybe because they never had a product schedule that counted on debugged refreshes), so Nvidia still find themselves in a difficult spot.

 

[demalion]

If I were an "old school" ATI engineer, I might get tired of everyone assuming we were incompetent before ArtX joine up. Looking at the R100 and R200, I see plenty of indication that the "old school" engineers are pretty good, and if I were to offer my own personal uninformed guess, I'd guess that ArtX primary contribution to the R300 was aiding in executing it on 0.15 micron successfully.

In lieu of any solid information to support my speculation, I will point out that Artx was spread out between both design teams, and the "ArtX Mojo" will be in the R400 too.

 

[Nagorak]

The R200 rocked, it's just it had immature drivers upon release... Part of the reason the R300 was so good is because ATi wasn't pressured to release it prematurely. Personally, I would have high hopes for the R400.

 

[Fuz]

I disagree, I believe its about the third iteration that they get it honed, so by thinking NV35 will be a tweak that will find itself hard against teh R350 and it'll be NV40 vs R400 that'll be the showdown of real heavywieghts.

You know, not too long ago, people were saying the "showdown of real heaveyweights" will be the R300 vs NV30. Its a shame that Nv didn't turn up for the fight.

The way things are headed, its going to be the R350 vs NV30, then R400 vs NV35. Of course this is assuming that R350 hits retail about a month after NV30, and there are no delays with the R400.

Being a product cylce behind is going to hurt for some time.

 

[overclocked]

Have someone yet figured out if R350 will use low-k?
And what "probably" minor improvments have they done?
12´t Mars right?

 

[Evildeus]

Actually nobody knows, and concerning NV35... well i think we should have some more information on NV30 before trying to found out IF NV35 will be a good chip

 

[overclocked]

Or we can just give a damn about ATI and Nvidia and lay your focus on when BitBoys will spank them both..

 

[nutball]

Where are Cirrus Logic when you need them?

 

[epicstruggle]

actually its going to all come down to the NV2100 vs R2100. Thats when all this will be decided once and for all.

 

[Mulciber]

actually its going to all come down to the NV2100 vs R2100. Thats when all this will be decided once and for all.

I think you're confusing it with the R2100 vs nv1500

if nvidias not shooting for 256mb of eDRAM with a 2048bit bus, it'll be another blunder!

 

[R300King!]

After looking back on this whole NV30 mess...I think Nvidia would have been better off just releasing their intended NV30 at their intended date (a few months ago and at the lower speed) and lose to ATi, than to lose face to their fan base ..and everyone else. They got whipped..and they won't admit it.