NVIDIA GF100 & Friends speculation

:idea: Maybe 768Mb framebuffer(and 192bits bus) is not enough. [for Heaven 1920x1080 8X]
Waiting tests for 1Gb version...:???:
 
L233 said:
In April last year I bought a GTX260 SP216 for €150. Now, 16 months later, nVidia's next generation mid-range card is quite a bit more expensive and, in most cases, barely any faster. What's the point?
Click to expand...

Well that really nasty price war that weakened both companies since the launch of Rv770 kind of skews the value proposition of cards following (both Nvidia and ATI). I don't think either company wants to re-engage on that type of cut-throat price war if they can avoid it.

We can just continue to hope that neither company starts ramping up to start launching with 600 USD cards again, with lower cards being upped to match.

Regards,
SB
 
fellix said:
MSI GTX460 768MB review
Click to expand...

This post contains the exact same slides that were linked on the previous page and actually quoted in the post previous to yours :) And also since all are marked "GBT confidential" I think it hardly qualifies as a review - it's more like a marketing material :p (not that Gigabyte have to lie one way or another since they sell GPUs from both vendors - it's just not a review in the sense that someone not working for an interested party has actually got the card and tested it)
 
Interesting ALU:TEX ratio :LOL:

So it appears to be 48 ALUs per Polymorph Engine with 8 TMUs.

Also, it's interesting to see that NVidia has two cards of different bus-width/ROP specification called GTX460 :rolleyes:

And, erm, wow, NVidia still can't produce a fully functioning GPU with all 8 Polymorph Engines working?
 
neliz said:
Thanks! that finally brings the GF104 TMU amount out in the open..

http://www.adrenaline.com.br/files/...ei/2010-07-07_nvidia_comparativogtx460480.jpg

Now, was the 64 TMUs on GF100 by design or not :D
Click to expand...

Hmm.. Ill throw a stone on your theory, sorry :p
Here it goes:

If we imagine that GF100 was 128 TMU's at the start, and then reduced to half that gives 64.
BUT GF104 is NOT half GF100, its more like 3/4. And with 64 TMUs. So... GF100 would have 80 TMUs just like GT200???
Nah, I dont align by that conspiracy theory. Seems more to me that GF104 is a revamped Fermi architecture then anything else :)
 
In some ways very similar to G80->G92.

Though I think there's a general expectation that GF100 will be "replaced" by GF102. Which is presumably not going to happen until 28nm is working. For NVidia.

So GF104 is sort of a preview of GF102 in that sense, presumably with the double-precision stuff deleted and the reclaimed space used for an extra SIMD.
 
Jawed said:
In some ways very similar to G80->G92.

Though I think there's a general expectation that GF100 will be "replaced" by GF102. Which is presumably not going to happen until 28nm is working. For NVidia.

So GF104 is sort of a preview of GF102 in that sense, presumably with the double-precision stuff deleted and the reclaimed space used for an extra SIMD.
Click to expand...
Why not G80 vs. G84?
The GF84 was a preview of the G92.
 
You must log in or register to reply here.

Similar threads

M
Interesting old GPU architecture
Replies
17
Views
3K
Ailuros
Ailuros
T
GPU Stability testing for Nvidia cards - PowerSweep
Replies
0
Views
449
T2098
T
J
Graphic card/CPU configuration for best memory graphics in Flight Simulator 98?
2
Replies
37
Views
5K
homerdog
homerdog
R
HDR settings uniformity is much needed
2 3 4
Replies
73
Views
9K
Cappuccino
Cappuccino
D
Speculation and Rumors: Nvidia Blackwell ...
45 46 47
Replies
928
Views
104K
Kaotik
Kaotik
Back
Top