NVIDIA Kepler speculation thread

I just want to see the guts behind TXAA already. Is it a post process is it a new form of AA is it an MSAA + post process... And I want to play with it it... *Wishes he could be a reviewer right now*
 
Some things I noticed:

He says "the fastest, most efficient GPU we've ever built," so that means GTX 680 > GTX 590 in performance?
The Boost appears to be dynamic—goes from default speed to ≈1.10 GHz according to the graph in the video. Average boost speed appears to be around the reported 1.059 GHz though.
 
He says "the fastest, most efficient GPU we've ever built," so that means GTX 680 > GTX 590 in performance?
I know it's been used to suite the currently speaking faction, but in my books, GTX 590 is not a GPU, it's a SKU (well, two at least) or a product.
 
I just want to see the guts behind TXAA already. Is it a post process is it a new form of AA is it an MSAA + post process... And I want to play with it it... *Wishes he could be a reviewer right now*

I was told it's based on old and new frames and generates lags, not best for FPS players.. dont have more info than this..
 
Huh? The HD4890 ran at 850Mhz. AMD has been up in that range for a long time.
What has HD4890 got to do with anything? And how the hell is 850MHz "in that range" when G80's hot clock is 1350MHz.

HD2900XT ran at 725MHz.

Before G80 NVidia was at parity.
 
Some things I noticed:

He says "the fastest, most efficient GPU we've ever built," so that means GTX 680 > GTX 590 in performance?
The Boost appears to be dynamic—goes from default speed to ≈1.10 GHz according to the graph in the video. Average boost speed appears to be around the reported 1.059 GHz though.

The graph should be wrong, according to specs it's supposed to be ~50MHz increase.

GTX590 isn't GPU, it's 2 GPU's.
 
Altering fundamental aspects in the micro-architecture (at SM level) within a generation is unlikely.
On the other hand, with such fat multiprocessors, I simply can't imagine what a 16 SMX Kepler part will look like. It would have very weird packaging - something like the GF104 die, but far larger.

GF100 and GF104 had different SMs…
 
Asus GTX680, used to see +XX% boost on Asus's boxes, no sign this time..
2029115ux6f6r6d6gwwj9hr6na.jpg

http://www.chiphell.com/thread-390462-1-1.html
 
A more interesting question is "why didn't NVidia do it this way from day 1, i.e. G80".

It took them that long to reverse engineer and then copy AMD's superior engineered parts? From the looks of things... :p

Cant wait for reviews! I should still be up for them...hopefully a(nother) leak soon as well.

If it's $499 will definitely cause some price movement. And at least this time Nvidia wont be caught short on RAM at the high end like they have been for ages.
 
I don't expect them to price the GTX680 at $499, but I'll be pleasantly surprised if it happens. I expect it to show up at $549 instead...
 
Yeah: http://www.shopblt.com/cgi-bin/shop/shop.cgi?action=enter&thispage=0110040015010_BPC1508P.shtml Would have thought that high resolution perf would force this one a little lower, but, we'll see. Would expect some "aggressive pricing" off the bat anyway....

Cool, hadn't seen that yet, but that's exactly what I would expect. I also expect the prices to shift around a bit in the next 30 to 45 days, but if anything, I only expect AMD to reduce prices and not NVIDIA.

Let's be blunt: NV can price it mostly wherever the want, and nobody will bitch. It's obvious what happens when AMD tries to do anything like that, even without any competition ;)
 
Back
Top