How Close to 500MHz has NV got?

Umm, I remember reading somewhere the R300 ran cooler the Albatron Geforce 4 Ti 4600.. err I think so anyhow. It would take some time to hunt down that quote though. If you have better info I would take that over some obscure review I read some time ago.

 

I don't think it pales in comparison considering what each offers. The R300 doesn't have a fancy heatsink, and the power requirement is not much (IIRC) greater than a 4600 (I actually consider that quite a feat for as many transistors as it has). Also, the size difference I think is a significant achievement for ATI (though it doesn't matter much, as, as you've said, the Ti 4600 fits within standards).

No kidding! It must suck to be a 3d graphics IHV if the nv30 is "outdated" now, or is even so when it is available. I don't think not soundlytrouncing the R300 as some seem to have been expecting makes the nv30 any worse in actuality.

 

http://images.ukgamer.com/previews/fx/demo_large.jpg

A different card; this one says "speed 210"...

From UKGamer.com

Edit: Decided to change the image to a link, because it's a bloomin' big picture...

 

OMFG it is getting faster by the hour.... My sakes man you should just post the link.

EDIT:Thanks...

 

Thats a good point, and imo it is a very important point too. If you have a look at the launch partners thats on the nvidia site, there aren't any "blue collar OEM's" at all.
No matter how you look at it, that can't be a good thing.

Edit: Having a look at that page again, it does list Fujitsu as an OEM, so thats one I guess.

 

Hmmmm.....that looks like the word "speedo", unless that's not an "o" after the "d".... Anyway, what does a swimsuit have to do with the GeforceFX?

 

Actually, if you look at the top right of the label (bottom right in the picture), there's some writing that could well say "500/500"...

 

I noticed that too, but to me it looked like "300/300". However, on further inspection I agree that it looks like "500/500".

 

Yep, 500/500.

Now we have to decipher the "Speed" label

 

What's with all this discussion about whether these alpha cards are running at 500 MHz or not. Is NVidia going to allow any benchmarks on them? If not, who cares... They could be running at 4.5 GHz for all it matters, if they don't benchmark them and production units are not available at that speed now, then when difference does it make?

 

I think the large (Understatement?) fan is probably required. I would think this would mean that nVidia's going to push the yields--and that they are anything but super for 500MHz clocking. Probably at 400MHz things would be entirely different--but I can't really say that for sure, of course.

Since the pricing seems to be divided, my guess would be that the $500 model will be the "Performance" model and carry the large fan (essentially an overclocked nv30) and the "cheaper" $400 model will be clocked 50-100MHz slower--and possibly *not* require the large fan, and thus be suitable for OEM (system OEMs) placement. The cheaper chip will probably run at the actual clock speed the chip yields for a somehwat cooler operation--and since that won't require the larger fan I'm guessing the clock difference might be closer to 100MHz than to 50Mhz.

Just speculating here--and I could be all wet if yields are poor all the way around right now--in that case they'd all need the super fans. But I don't think it would make sense for nVidia to do a non-system OEM-useable version of the chip and nothing else. So I'll bet the lower priced models will run ~20% slower and not need the gargantuan fan. What say you?

 

It also says 'failed' below it, plus something else that's completely undecipherable.

 

damn i forgot what i was going to say now....

shit, i give up.

 

I wonder if it's possible that XBox 2's GPU (assuming its from Nv) could reach 1 Ghz


....if they're going to hit 500 Mhz in Dec 2002 or Feb 2003

 

It also says "Chip # 7391" (or maybe "Chip # 739.1", which makes even less sense) in the upper left corner. I find it very hard to believe they have over 7000 NV30 chips mounted on cards and tested. Maybe the 7391 is the Wafer Lot number, and the #210 is actually the chip number designator from a particular batch of wafers. (Just idle speculation)

 

Walt: Interesting speculation. I assume you're referring to the "Ultra" and Vanilla versions? Because I'm under the impression that both version will require the huge cooling system, and that they will later announce a single-slot version with lower clock speeds.

 

A $400 graphics card is not suitable for OEMs anyway. Remember these are the guys who sell P4 chips coupled with TNT2s.

 

Well IFF the pics of the wafers I saw were accurate then it looks like a wafer holds about 125ish parts. Again thats assuming those pics were accurate.....

 

I'm suprised no one actually touched on the naming scheme of the chip first.Nvidia gets the no. of there chips by the generation followed by the memory speed so if it is 5800 for instance this is there 5th generation Geforce part since there doesn't seem to be anything new at all & with the 800 being the Mhz of the memory speed.However if they do both core & mem. asynchronously we're looking at 400 Mhz for the core.




Now the Ultra version might be explainable in that it just might feature a 256 bit memory interface even though current reports currently blow any theory of such a feature being supported.However,it definitely explains the the same 5800 no.


ALL imo but still a very sound theory.