Crytek held a Crysis 3 Community Day, in which Sean Tracy showcased some nice and new graphical features of CryEngine 3 - that will also be supported by the upcoming Crysis 3 game. ... CryEngine 3 Presentation at Crytek
Really impressive stuff, if i had a 690 i'd be sitting rather pleased with the world in general right now, just being honest i just wish my 680 was gonna be enough for this, i don't use any AA, blur nor Post Processing and it'll still destroy my £500 card, i mean lets take a step back here folks and have another look, this game is gonna need around £750 of gpu power to run 1080 60fps Ultra.
These guys are as bad as DICE, moar lens flare... @Texter - It will probably ned two to three of those hahaha, remember Crysis was supposed to run smooth maxed out on an 8800GTX.
Just came off from the Beta, it doesn't really require super strong computers. I ran all on max settings but reduced the AA and stuff down to the lowest. Game does look really nice for a change.
Such a card never existed last year. Unless you’re meaning some sort of GK110 card, but that was clearly not ready in 2012, and in 2013 will be around £800 not £500. This engine will perform as well as you want it to perform. If your one of the lucky ones who dislikes post processing, motion blur, AA and lens flare then you won't need much hardware at all as those things can all be disabled, even then the I’m sure you will still be able to get 60fps if you fiddle with some settings, and probably not even notice much, if any difference. Maxing it at 60/120fps just for the sake of maxing it is a different story, but those people tend to be hardware enthusiasts so will already have the best GPU/CPU combo going. Unfortunately post processing, lens flare, AA etc are all very important to me and I won’t be turning any of those off, so will need to be a new GPU for me, but i always skip a generation so I’m due one anyway
The number of people who still believe the rumor that GK110 was somehow intended to be the 6xx series is insane. Moving from 40nm to 28nm and increasing the die size by like, double, on a new architecture, yeah, that was their move, sure guys. Maybe if they wanted to see how quick a yield-less fab could go out of business. For some perspective, the die of the presumed 780 (current K20) has more transistors than a 690. No way they were going to try to do all of that in a single generation.