Originally Posted by Booty Warrior
Out of curiosity, has anyone really tried gaming with that much AA? I personally don't notice any IQ improvements beyond 8xQ.
Meh, I don't see too much on the horizon that's really going to push the hardware boundaries. Crysis 2 is shaping up to be a walk in the park. DA2 will be cake once the drivers get sorted out (it already is on the red side).
Maybe BF3? It seems to me that "enthusiast" hardware is rapidly outpacing game development while devs shift more and more focus on consoles. For all intents and purposes Fermi setups can max everything on the market atm... yes, some games require more than one
, but still...
I wonder how many people will feel the urge to upgrade to Kepler if games still don't even take full advantage of the current hardware. 3D surround users maybe?
That is exactly my point.
I did not mean that Fermi was not fit for future gaming requirements.
Just that its main purpose and focus right now, is not to elevate the the performance level or really deliver a more evolved gaming experience due to innovations like DX11 ,tessellation etc that are very premature at this point,than rather to sell and provide Nvidia with sufficient funding to engange the study and development of 28nm architcture at its peak.
Maxing out a game is not always a factor of true performance value.
To make an example....GTX 295 Quad SLI will equally max everything just like a 580 but still the gaming experiences differs because of the far evolved architecture of 580 compared to the 295 and that is two gens skipped
The same experiences will be maid with 28nm just that it will feel more suited and far more superior even though the same FPS are reached..get it?^^
And the most relevant factor to me which I believe most of you will probably not consider is power consumption.
28 nm is said to perform 3-4 times more efficient than fermi in its power per watt ratio.
To me not only the fact that Kepler will draw much less power to achieve triple the performance Fermi will is an interesting point rather than the how much cooler such a chip must theoretically run.
The Jump from generation to generation always brings new innovations.
But the actual performance differences were never that great between each gen.
But if you put Kepler into that list the actual evolvement of the chip and performance increase will not only be much more obvious it will be devestating
Also after Kepler there will be blank for 3 years at least according to the road map^^ that is the bluntest but shortest aspect of the chipxD
Originally Posted by Kung Pow
This is OCN.net
The question shouldnÂ´t be if a card is relevant or overkill rather than how far can it be overclocked to be able to achieve the successors performance level.
Also all this DX11 ,tessellation etc hype is in my oppinion more advertisement than reality right now.
A statement has been made a few pages back that"580,590 and 6990 might be overkill now but in 6-8 months they will be able to be fully used".
I personally think that this is not there purpose.
By the time tessellation and DX11 have reached their peak in development and games are fully capable of ultimately taking advantage out of them a new branch of graphic chips will arrive which are not only destined and specified for the performance requirements but in one word just evolved.
All this hype is just the end of a generation of gpus, I am sure 28 nm will bring innovations and be more suited for whatÂ´s to come.