as long as nv can handle the pressure from the console effect for 1 more year, nv will have nothing to worry about. a new design is about 3 years? we are already more than 2/3 of the ways through the lack of support for some of the dx12 features period. by late 2017 or early 2018, nv would have an answer for asynch compute or other dx12/console features for sure. with how well their marketing is, and full spread of sponsored games, I don't foresee any problems for nv at all. it is pretty incredible how well nv handled dx12 and the console effects so far.
amd couldn't capitalize on their advantage shows how bad the top management at amd is.
as a gamer, I wish amd luck. they are the only thing standing in the way of 500$ mid range gpus.
People said Pascal was supposed to be their answer for Async compute, and they're still bad at it.
This is going way off topic. I was just answering his comment about nvidia's supposedly bad business decisions. I never said anything about gamers being better off.
It doesn't matter at the high end when the 1070 and 1080 are beating the Fury X in DX12. We'll see how things pan out with 1060 and 480 - I think it'll be interesting to see, as it is more important the midrange and lower segments.
However, even though Vega 10 will probably beat the 1080 in DX12, it'll probably lose to the Titan P or 1080ti. AMD are too early with their dx12 advantage, and too late with their high-end lineup (Vega). Once DX12 is important and has more than a handful of games, we'll see nvidia having full dx12 cards thanks to their domination and profits from DX11, and I'm not sure if AMD will keep the advantage they have today.
I do agree though that reviewers are not highlighting AMD's current DX12 advantage at all, except maybe Hardware Canucks.
You're comparing two different generations. The fury x destroys the 980 ti in async games. Even the 390 gives it trouble, and that's a mid range vs a flagship.
They have at least four cards left. A 490 will likely beat the 1070. A 490x will likely beat the 1080. Then, they have a vega-Fury and a vega-Fury X. This time, they won't have to wait a year for devs to take advantage of their hardware since dx12 and async are becoming the norm.
The console effect and async compute may be an advantage for amd, but will it be enough to compensate for the 50% higher clock Pascal has compared to Polaris?
We'll see. As of now, Nvidia is not performing very well in async games. Maxwell is terrible at it, and even Pascal isn't seeing as big of gains as I thought it would considering the 1080 is a $700 card (no one sells it at the claimed $600 msrp).
The only thing they have left is a Pascal Titan and a 1080ti (x80ti cards are usually just cheaper titans more or less).
I was on the Pascal hype train early on until I've seen benches that show that it doesn't gain much from dx12 AND tweaktown's benches showing that two 480s can beat it.