NVIDIA GF100 & Friends speculation

Usually NV time - Pacific

 

I'm so confused by the angst over Fermi. Isn't cool new tech simply cool new tech? These aren't football teams. They're fun-enablers and any new tech is good tech. Can't wait for real data.

 

+1

pushing tech is always good and competition between graphic card producers is just a win-win situation for us consumers.

 

if some rumors I heard "behind scenes" is correct, this thing is going to perform amazing in DX11 tessellation mode of Unigine.

 

So assuming the Fermi can do 2 triangles/clock...what does this mean for real games like crysis? Would it be possible to double polygon counts without any hit in performance? Crysis would look a LOT better if things could be smoothed out a bit.

 

You dont need "behind scenes" rumors, Razor has said in this very thread that Fermi is upwards of 50% faster than Cypress in that benchmark.

 

and 220% faster than the GTX285 in games and demos. I can't wait, we need a good ol price war.

 

hah, I don't think that's likely to happen - except in a minor way.

as has been pointed out here before, AMD needs profit desparately, and nvidia's manufacturing costs for the Fermi look to be very high. It will likely be in neither's interest to engage in anything more than a little price jiggling.

 

With 8xAA there can be quite an impressive performance jump over GTX285.

 

Question about benchmarking tesselation, part of tesselation pipeline is vendor specific, if I understood correcltly it is the fixed function part, couldn't this part vary in quality from one too other too much making the comparison of fps pointless? And couldn't a vendor generate a lighter or heavier load based on "free" resources on the chip and so always keep the fps close to, let's say, 30 fps while other vendor just bypass?

 

there is no fixed function portion to tesselation when it comes to coding or hardware. The code just runs better on a given hardware because of the hardware advantage.

 

Since Fermi is supposed to support C++, I'd love to see a benchmark of the open source Stockfish chess engine. If anyone happens to get their hands on one...

 

From PDX LAN 2010:


More Photos: http://www.facebook.com/photo.php?p...p?pid=3269565&id=8409118252&fbid=252645433252

 

So... why choose 3DM Vantage Performance score? Hey, it's an enthusiast part, it HAS to be using the eXtreme preset...

Could PhysX be needed to show any performance advantage over GTX295?

Could higher res/shaders complexity lead to lower performance than the competition?

 

Yes and its quite a bit

No, and there is quite a bit of an advantage to Fermi when it comes to shader complexiity.

 

X score with those P22K is 15K, as I said on XS, not sure if those numbers are legit.
I assume they're fake but show up at DH a couple of hours later..

edit: DVI single link nFinity, aftertought or patchjob?

 

Poor Tom Peterson. Vegas to PDX LAN. He looks tired.

 

Probably can't transfer too much data over SLI either- perhaps one reason why CF doesn't allow Eyefinity 6.

 

So, why choose P score if X would give it the lead anyway?

Something doesn't add up.

 

Single Link? Did you know that they need twice the bandbwith with 3D Stereo and 120Hz?