NVIDIA to release 10 new cards before NV30?

[Pav_37]

I dunno, if this is true, I wonder who will actually buy these? lol - I mean, they just release them too much, don't you think? Best work on better and faster drivers first

Source: http://www.warp2search.net/article.php?sid=5949&mode=thread&order=0

 

[Brent]

they are scared, scared they are gonna loose market and money with the R300

so they are pullin out all they can to get NV28 out i bet, try and give the impression they are still in the game...

though, only adding AGP 8X aint gonna do it, no sir, nope

if all they did was truely just add 8x agp support then IMO that's just stupid

they better be clocked at 400 core and 750 memory or else, hah

i dunno, just seems kinda lame to me if that's all they changed, too many cards for sure

[edit]
though at the same time i guess it would drive GF4 prices down perhaps...

 

[Pav_37]

well there's no detailed specs so we can only hope and assume that they did make them faster, but they would have to be a lot faster to make people buy them, unless NV30 has been delayed untill early 2003.

 

[RussSchultz]

I don't want to start another NVIDIA v. ATI slugfest, but complex ASIC designs take more than a few weeks to pull off.

These are the designs that NVIDIA came up with months ago, and it has nothing to do with any recent product releases of ATI, or whether or not they're scared.

I do agree, however, if AGP8 is the only thing added, somebody should fire whoever gave those products the green light.

 

[Dave Baumann]

Adding AGP8X to GF4 without substancial changes won't be so bad (although, the slides shown before indicates that NV28 may have a few more changes than just that I wonder if they've increased the number of vertex shaders or something), its whether or not that added anything more to to NV18 than just AGP8X - I think it will be a real mistake in the light of Trident, SiS and ATi's new low end parts not to have a low end DX8 part.

 

[RussSchultz]

Die shrinking the 42/44/4600 would make it a cost competetive DX8 part (i.e. on the likes of the Radeon 9000)

 

[Randell]

Die shrinking the 42/44/4600 would make it a cost competetive DX8 part (i.e. on the likes of the Radeon 9000)

if that is the NV18 in effect, then that's an excellent response. Can we assume it was na early Nv28 that nVidia tried to roun Doom3 with at E3 or is that such a stretch?

Remember it took over 9 months for a mid-cost DX8 (gf3Ti200) part from NVidia, so maybe their plan was to have 'really' capable DX8 parts in the low end and mid-stream and only have one top end DX9 part until Q1/2 03? After all ATI still dont have the $200 segment covered by a DX9 part either until the 9500 is out and we dont know how that will perform.

 

[RussSchultz]

I would guess the NV28 would be the die shrunk 42/44/4600 and The NV18 would be more related to the GF4/MX line.

But I probably have the madness of their product numbering schemes a bit confused.

 

[Mulciber]

I would guess the NV28 would be the die shrunk 42/44/4600 and The NV18 would be more related to the GF4/MX line.

But I probably have the madness of their product numbering schemes a bit confused.

I believe that Jensen said in a conference call last month that the nv30 is the only chip being designed on .13u right now. So nv18 and nv28 are slated for .15u. The transistor count rumored on both is extremely high though, the nv28 is supposedly 81M transistors. It will take a lot of oomph for them to be able to market that card as a viable mid-range card though if the ATI 9500 speculation is worth anything.

 

[Brent]

i guess we'll just have to wait for specs to know for sure if and what has been changed..

 

[McElvis]

As far as I remember from those slides, that 8 was actually a 6.

The text was very hard to read.

 

[BoardBonobo]

nVidias PR department will be out in force with magic markers any day now

 

[Hellbinder]

It really bothers me that Nvidia is going to release a GF4 Ti ultra or some nonsense making it the 3rd product in a year released with virtually the same core technologies.....Isn't this the same thing that 3dfx was doing in the end??? releasing the same tired technology over and over with little change.. while the rest of the industry moved on?

What bothers me is all the Nvidia sites, and 95% of the rest of the sites on the net are going to start praising the name of Nvidia Just because they released *another* rehashed product with a higher clock speed....Furthermore, what about all the vendors... They are going to have to clea their exsisting Nv25 stcok out to make room for the Nv28... just to make room for the Nv30 in a coupple more months.. Maybe Nvidia will just relrease the Nv28 and put the Nv30 on hold till January/Feb...

Ive no doubt that there will be many a review that state "buy the Nv28 now because it offers good performance and is Cheaper than the 9700.. and Wait for the Nv30 which will *clearly* be faster and better"

 

[jb]

Yea we all know is not an easy thing to be a nVidia board vendor.

 

[LittlePenny]

HellBinder, I certainly don't see the problem with an nVidia fan site recommending NV based products. As the same goes for ATi and their products.

But as for the other sites praising NV for a 6 month rehash cycle, well some justification should be given. It certainly isn't an easy or cheap process working and refabbing on a frequent basis. As for saying they deliver a new architecture more quickly than ATi, well, that is just disinformation.

 

[Pav_37]

Well if NVIDIA is to stick to this 6months schedule then we will see announcement of NV18 and NV28 this month since August 6th is exactly 6 months since Geforce4 was officially announced. Where does that leave NV30 though? February again?

 

[Hellbinder]

Its the nature of the beast..

I just wish Conan was around to deal the beast a death blow... Its time for people to recognize endless rehashes of the same product for what they are.. and not worthy of 400$ price tags, and the Praise of millions.

ATi has stated recently that with two full time Design teams we can expect them to have a 6 month product cycle as well.. So it looks likte Everyone is going to start doing it

 

[Pav_37]

haha, damn, guess we gonna neeed to save money all year around to get all these new cards when they come out

 

[Galilee]

But what if NV28 comes out by the time R300 hits the market:
Lets imagine:
- NV28 as fast as R300
- Faster FSAA than GF4
- More shaders and stuff making it faster in "new" games.
- 100$ -150$ cheaper than R300. Maybe a little less than what Ti4600 is now.
- Dx8 product

That would make it a hell of a nice card, even if not supporting dx9.

 

[multigl2]

IMHO the NV28 probably wont be as fast as the R300; and as stated above, the slides were confusing and blurry, the 8 could be a 6, so 61 million transistors, not 81. that would be 2 million less than the current GF4 has...