nvidia "D8E" High End solution, what can we expect in 2008?

So what can we expect in fall 2008 then ? Geforce 10800 using the GT100 or GT200 etc. silicon ?

GeForce (10) D880. :LOL:

Do we know how well the 9800GTX is going to perform relative to the 8800GTX?

Should be nearly the same performance overall, in ALU or Tex depended situation higher, since it has 64TMUs@~800MHz(1.4x G80GTX) and 128ALUs@~2GHz (1.5x G80GTX), problem is BW, but here may G92 better pixel-compression may help.

btw.
Originally G92 was planed to reach 2.5GHz @ ALUs, so we may see some nice AIB-OC-versions...
 
Last edited by a moderator:
GeForce (10) D880. :LOL:



Should be nearly the same performance overall, in ALU or Tex depended situation higher, since it has 64TMUs@~800MHz(1.4x G80GTX) and 128ALUs@~2GHz (1.5x G80GTX), problem is BW, but here may G92 better pixel-compression may help.

btw.
Originally G92 was planed to reach 2.5GHz @ ALUs, so we may see some nice AIB-OC-versions...

Hmmm RV770 might actually be enough to take the performance crown then :) Should be interesting. Btw whats with the D8E naming convention? is this a b3d invention or is this is actual code name being used internally @ NV?
 
Hmmm RV770 might actually be enough to take the performance crown then :) Should be interesting. Btw whats with the D8E naming convention? is this a b3d invention or is this is actual code name being used internally @ NV?

Don't remember where the DxY codenames came from, but I got a feeling that they're in fact real codenames, but not for chips, but rather cards.
 
I'm sorry, but you're wrong.
The 8800 GTS 512MB (G92-based) already successfully competes against the 8800 Ultra, so, anything with the same GPU on a higher core, shader and memory clockspeed should be even faster.

Not with AA applied. These specs for the 9800GTX are poor. It's basically a GTS with slightly faster memory. In which case it will still lose to the 15 month old GTX at high resolutions + AA. Buying a GTX at launch really would've been one of the best purchases ever.

http://www.computerbase.de/artikel/...e_9600_gt_sli/24/#abschnitt_performancerating
 
Not with AA applied. These specs for the 9800GTX are poor. It's basically a GTS with slightly faster memory. In which case it will still lose to the 15 month old GTX at high resolutions + AA. Buying a GTX at launch really would've been one of the best purchases ever.

http://www.computerbase.de/artikel/...e_9600_gt_sli/24/#abschnitt_performancerating


It's not like there's any doubt that this "9800 GTX" will be cheaper than the 8800 Ultra/GTX, and that the "9800 GX2" will slot in the true high end market until the Geforce 10 arrives later this year. ;)
 
Well price is one thing but I think Domell was referring to performance. We also already have a good idea of what a 9800GX2 would do by looking at 8800GT SLI numbers.

http://www.firingsquad.com/hardware/geforce_9600_gt_sli_performance/page8.asp

Still not very impressive IMO.....

The 8800 GT and 9600 GT can't do Tri-SLI or Quad-SLI like the 9800 GTX/9800 GX2 can. ;)
Still, i think no one was expecting this "refresh" to be a huge jump anyway.

The 7900 GTX and 7950 GX2 weren't that much faster than the 7800 GTX 512MB in real world apps either.

Nvidia does seem to be milking the G92 core to the very last drop, though (8800 GS, 8800 GT, 8800 GTS 512MB, 9800 GTX, 9800 GX2, 8800M GTS, 8800M GTX, QuadroFX 3600M, QuadroFX 3700/3500, etc, etc).
 
Last edited by a moderator:
Why did Nvidia take this odd SKU approach with the G92? Why not release a 8900 GT, GTS, GTX & GX2 as one product line, within a reasonable amount of time and call it a refresh? As it is the GPU isn't capable of much more than the GTS is providing so any other single chip SKU they push out has no choice but to underwhelm.
 
Why did Nvidia take this odd SKU approach with the G92? Why not release a 8900 GT, GTS, GTX & GX2 as one product line, within a reasonable amount of time and call it a refresh? As it is the GPU isn't capable of much more than the GTS is providing so any other single chip SKU they push out has no choice but to underwhelm.

It's a combination of factors:

AMD has no single-GPU SKU that can touch the G92 too in pure performance (much like what happened with G80).

But, on the other hand, it's understandable that Nvidia would want the G92 not to completely overwhelm any G80-based 8800 GTX/Ultra, based on the fact that AMD doesn't seem to be interested in pursuing the high-end single-GPU crown either -primarily for cost reasons-.
Also, this approach will likely benefit the hypothetical marketing impact that the Geforce 10 performance will bring later on.
 
it's understandable that Nvidia would want the G92 not to completely overwhelm any G80-based 8800 GTX/Ultra
Or maybe they want, but they can't because there's no fast enough memory? =)
Does G92 support GDDR4?
What if that rumoured new G92 revision (not that i believe that there is one but...) will support GDDR5?
From G94 benchmarks it's pretty clear to me that G92 is heavily B/W limited in most cases and what they really need right now is not higher core clocks but faster memory...
 
Or maybe they want, but they can't because there's no fast enough memory? =)
Does G92 support GDDR4?
What if that rumoured new G92 revision (not that i believe that there is one but...) will support GDDR5?
From G94 benchmarks it's pretty clear to me that G92 is heavily B/W limited in most cases and what they really need right now is not higher core clocks but faster memory...


If that was really the case, a 0.8ns GDDR3-based G92 card would have surfaced long ago...
 
Some GF9800GTX specs? :) Then we can say GF9800GTX<GF8800GTX.
http://en.expreview.com/?p=274

Such an SKU will suffer in high resolutions against a 8800GTX due to less VRAM; if those specs should be real, I'd love to know what the 9800GT will look like. Current rumours point at 9800GX2 & 9800GTX being meant for the high end segment, 9800GT Performance and 9600GT for mainstream. Those specs sound fine for a 8800GT replacement. The 9600GT is already damn close to 8800GT performance.

Degustator,

Does G92 support GDDR4?

Either way the better question would be if NVIDIA is even willing to touch GDDR4 at all.
 
Well price is one thing but I think Domell was referring to performance. We also already have a good idea of what a 9800GX2 would do by looking at 8800GT SLI numbers.

http://www.firingsquad.com/hardware/geforce_9600_gt_sli_performance/page8.asp

Still not very impressive IMO.....

Especially if you take into consideration rumors that Nvidia had to lower the clocks for each core on the 9800GX2 to make a single card solution. And that said issue was a primary motivator for delaying the lauch of the card.

Now, whether that actually was the case or not, who knows.

Regads,
SB
 
Such an SKU will suffer in high resolutions against a 8800GTX due to less VRAM; if those specs should be real, I'd love to know what the 9800GT will look like. Current rumours point at 9800GX2 & 9800GTX being meant for the high end segment, 9800GT Performance and 9600GT for mainstream. Those specs sound fine for a 8800GT replacement. The 9600GT is already damn close to 8800GT performance.

GF9800GT specs :(

9800GT: 112SPs, 650/1625/1000MHz.

Well, if true then i ask what is going on with NVIDIA? It seems there will be NO performance boost over GF8800GTX. Moreover it seems GF9800GTX will be SLOWER than GF8800Ultra/GTX in FullHD resolution or 1680x1050 with AA enabled due to only 256-bit memory bus and 16ROPs.
GF9800GT is believed to be slower than current new GF8800GTS too according to specs :(

It`s first time in GPU history when forthcoming highend GPU ISN`T visibly faster than it`s precedessor :( Could we have any hope that G100/GT200 will recompense us all this "GF9" situation?
 
Last edited by a moderator:
GF9800GT specs :(

9800GT: 112SPs, 650/1625/1000MHz.

Well, if true then i ask what is going on with NVIDIA? It seems there will be NO performance boost over GF8800GTX. Moreover it seems GF9800GTX will be SLOWER than GF8800Ultra/GTX in FullHD resolution or 1680x1050 with AA enabled due to only 256-bit memory bus and 16ROPs.
GF9800GT is believed to be slower than current new GF8800GTS too according to specs :(

It`s first time in GPU history when forthcoming highend GPU ISN`T visibly faster than it`s precedessor :( Could we have any hope that G100/GT200 will recompense us all this "GF9" situation?

That's one possible interpretation. Pop quiz: what's the other interpretation?
 
Back
Top