NV50 SPECS ???

M

Martin Eddy

Regular
Just saw this on WARP2SEARCH, apparently it came from THG.

I've come accross some interesting new details on the nv50. remember it can be just a rumor or dunno, so take it with a little grain of slat :)))

These are the Specs for Nvidia's NV50

# 375-400 million of transistor on 90-nm process
# 550-600 MHz core
# 1600 MHz 256Mb (Support for upto 512Mb) GDDR3
# Graphics Core 256-bit
# Memory Interface 256-bit
# 32 (24 real) pixels pipelines (on the Ultra model)
# 32 Vertex Shader Engines
# 51.2 GB/sec Bandwidth (GDDR 3)
# Fillrate 12.8 billion texetls /sec.
# Textures per Pixel 32 (maximum in a single redering pass)
# Nvidia's LMA - IV
# DirectX 9.0d UMI-23
# Release date Q3 2005.
Click to expand...

Sounds as unrealistic as the R520 specs goin' round.

Any thoughts from the experts here?
 
I believe the wafers these chips will be fabbed on are condensed, hardened and then sliced from the stuff dreams are made out of. ;)
 
ANova said:
More than likely just another fan's dreams in response to the R520 rumors that are going around.
Click to expand...
Maybe radar?
But then again there are probably thousands of other potential people that would do this
 
If the NV50 exists, as in its finished, the number of transistors should be known, surely? Why would the transistor count be a guess, if this info was concrete?
 
There is nothinng wrong with most iof the r520 specs "floating around". In fact i just dont see what is so funny. Please point out what is so "Pie in the sky" about a SM3 part with 24 pipelines???

As for the Nv50 its not a spring release product is it?
 
It will be interesting to see if NV is significantly later to market with a 90nm top-end card than ATI. Last major transition, as we all recall, there was a lot of pointing at the difficulties of making such a move for the lateness to market of NV30. There was also a good bit of grumbling in some corners that ATI had somehow temporarily cheated by sticking with 150nm for R300, but that they'd eventually meet their Waterloo with 130nm a little later down the line. So far as I can tell, this never actually happened.

And NV's CEO, in the recent conference call, very clearly (in my mind at least), refused a softball opportunity from one of the questioners that was clearly hoping he'd say that the 90nm transition was not going to be as difficult as 130nm. Jen-Hsun just wasn't gonna say it.

So if ATI beats NV to market by a significant margin (say 90 days+) at 90nm top-end, I wonder if we might be getting to a place where it would be legitimate to say that ATI is just a little bit better at these things. . .

Or maybe they won't. "1H" and "3Q" internet rumors (respectively) are a little thin to hang our hats on just yet.
 
51.2GB/s is definitely not enough, think about higher resolution, HDR, multi-pass rendering, better AA/AF. But if it has a big on-die cache, which can hold most of the intermediate result, thats another story(again, this is not quite possible, given the listed transistor count).
 
991060 said:
But if it has a big on-die cache, which can hold most of the intermediate result, thats another story(again, this is not quite possible, given the listed transistor count).
Click to expand...
If you can't guarantee most of your intermediate results fit in this big cache..well..that cache (or local memory..) will not buy you (almost) nothing, performance wise.
 
digitalwanderer said:
I calls BS....I still think Q3 is too soon for the nV50.
Click to expand...

If they are going to be as late in the top-end as you suggest, it will be interesting to see if they try ATI's strategy from last time and release a low-end card in the interim at 90nm to gain some experience. Any disturbance in the Force along those lines? Say a 6600 at 90nm with a tweak or two? S/b higher yields and clocks. . .
 
Maybe Q3 coming from NVidia is a bit more realistic giving that ATI's current lineup of cards are about easy to find as a virgin in Vegas.
Or maybe I should say they can be found but no where near at the MSRP that they say ;)
Seems both companies are guilty of announcing unrealisitc dates as to when their cards will be available...or at least available to the public in any quantity.
 
16Valve said:
Maybe Q3 coming from NVidia is a bit more realistic giving that ATI's current lineup of cards are about easy to find as a virgin in Vegas.
Or maybe I should say they can be found but no where near at the MSRP that they say ;)
Seems both companies are guilty of announcing unrealisitc dates as to when their cards will be available...or at least available to the public in any quantity.
Click to expand...

It does seem to me that one of three things is likely to be true. Either the ATI dates that are being bandied around are too soon. The NV dates are too late. Or we will see an interim "confidence builder" (both for internal and external audiences!) 90nm NV part.

ATI does have more impetus to get out of the gate soon, imho. They want to come to the SM3.0 party. They always said they did (however it's been spun by others). They've just said they didn't want a "checkbox" --they wanted performance-useful. I'm sure their marketing folks want the checkbox as soon as possible, and I feel pretty confident that their technical folks are feeling a bit unfairly picked-on on this point over the last year and have a glint in their eye to bring the high hard one as soon as possible to validate what they've said over the last year, for professional pride and competitiveness.
 
madmartyau said:
Just saw this on WARP2SEARCH, apparently it came from THG.

I've come accross some interesting new details on the nv50. remember it can be just a rumor or dunno, so take it with a little grain of slat :)))

These are the Specs for Nvidia's NV50

# 375-400 million of transistor on 90-nm process
# 550-600 MHz core
# 1600 MHz 256Mb (Support for upto 512Mb) GDDR3
# Graphics Core 256-bit
# Memory Interface 256-bit
# 32 (24 real) pixels pipelines (on the Ultra model)
# 32 Vertex Shader Engines
# 51.2 GB/sec Bandwidth (GDDR 3)
# Fillrate 12.8 billion texetls /sec.
# Textures per Pixel 32 (maximum in a single redering pass)
# Nvidia's LMA - IV
# DirectX 9.0d UMI-23
# Release date Q3 2005.
Click to expand...

Sounds as unrealistic as the R520 specs goin' round.

Any thoughts from the experts here?
Click to expand...

Isn't 90nm half the size of 130nm? If that is true then it is possible to have 32 pipelines.
 
You must log in or register to reply here.

Similar threads

S
  • Locked
Final specs for RSX and Xenos - please help me to found them
Replies
3
Views
5K
StefanS
S
E
Updated *fake* Rev specs?
2 3
Replies
46
Views
8K
Teasy
T
Back
Top