Do you think Mantle is going to catch on? The way I look at it, if it was only available to AMD cards, devs wouldn't waste their time trying to develop for it (Aside from those partnered with AMD). If it is available to NVIDIA cards with a patch (which is what I've heard is the case), couldn't NVIDIA effectively destroy it just by not making their cards run with Mantle (bringing about the first scenario where devs won't bother with Mantle if it's only on AMD cards)? I suspect there would be some sort of optimization with Mantle to work better on cards with GCN architecture so AMD gets a little boost in performance compared to NVIDIA cards. Would NVIDIA actually do it, since their performance would be better than in DX11, but with Mantle they would be lower performing than the competitor, or would they not bother making their cards run with Mantle so that it doesn't catch on?
If mantle does end up running on NVIDIA cards, I'd probably end up grabbing a GTX 770 for my next card (unless 800 is out by then) so I can take advantage of Mantle and of Shadowplay (and for some reason NV actually has better price to performance in Canada... GTX 770 is cheaper than R9 280X/7970) =P unless AMD has plans to make their own tool similar to Shadowplay? Though that might be irrelevant seeing as how YouTube in 2014 will apparently be going super hardcore on copyright and gameplay might not even be allowed on YT, and if it is it will be going through like a 1 week approval period X__X
Mantle is designed to run with any GPU that uses the GCN architecture. So any AMD GPU product.
As the new generation of consoles (PS4, Xbone and Wii U) all use AMD APU's there is a huge market right there with the capability of using Mantle. If dev's use Mantle, it will dramatically reduce the time required to port a game between consoles and PC. AMD is currently sat in a very favourable position for Mantle.
The more sensible question, IMO, would be if Nvidia were to make a similar programming layer, would it be as widely adopted, considering there market is mainly in PC and tablets.
Nvidia doesn't really lose anything by adopting Mantle, but I don't know that they will. They might be too stubborn to say "hey AMD invented this awesome thing so we're gonna support it now." I think it's more likely they will bring out their own similar API and try to compete.
Hopefully this doesn't turn out like when AMD invented the 64-bit architecture we're all using in our CPUs these days, where they basically got no credit for it and few people today recognize that AMD has been any kind of leader in the CPU space.
The API being used in consoles is NOT mantle. Consoles use DIRECT hardware level access APIs. Mantle is a Low level API, not direct. The code is more similar to the consoles API than DirectX or OpenGL, though.
This could either catch on, and become a New, universally used API (AMD even said that nVidia could adopt this if they wanted to) OR, It will become something that only a few video game designers use, and it will be trivial, like PHysX has become.
Everyone thought PhysX would be this huge game changer, and give cards a HUGE boost in gamin, and whatnot. But look at it. No body even uses PhysX anymore.
I think Mantle will catch on. It provides two main benefits which are basically unrivalled. First of all, it will supposedly make console ports a lot easier to do and more time can be spent optimising it for PC (I don't really know why, I just have seen several sources saying this). The second, and more promising reason, is the higher performance overall. It allows developers to directly access the GPU, meaning it can be used to the fullest extent using all of the GPU towards the game. This will especially increase FPS on very graphically demanding games, such as Crysis 3 should they adopt it (I hope they do!).
I want Mantle to be successful, but it cannot be implemented in every game. AMD have said that Mantle isn't right for everyone. So it depends how widely used it becomes. I wouldn't expect it to work in any serious RTS game.
AMD have come out and said that Mantle may not be suitable for every game. The mathematics needed for RTS games is a little different from many other games. And besides that, many RTS games don't focus on performance. They focus on a complex engine that can has to perform a high number of calculations.
Put simply, Mantle won't enter the entire library of PC games. Only games like BF4 where performance counts above much else.
Mantle is not about performance, it is about flexibility. In the beginning, that means most devs will use it for squeezing more performance out of the hardware. Later on, it could mean the emergence of new paradigms or new graphical effects which can't be (efficiently) done on Direct3D.
There is little reason to believe Mantle's usefulness is influenced at all by genre boundaries. In fact, if RTS games have unique requirements or priorities, they stand to gain the most from Mantle.
Total War games come to mind. The Total War engine has a unique graphical task in that there are thousands of soldiers on screen at any one time, hundreds or thousands of trees, grass all over the place, every projectile is rendered, at least when you're close, etc. Because this graphical situation is fairly unique, it's safe to say modern 3D APIs are not well designed to handle that much drawing in one frame. The engine can gain immensely just from being able to parallelize the rendering process, one of the main selling points of Mantle, let alone a custom rendering pipeline designed specifically for their needs.
Actually, the more different a is game from "the norm" -> Direct3D is less built for its needs -> more potential benefit from Mantle. Battlefield 4 could actually be one of the worst examples of what Mantle can do (being a moderately-sized, lightly-populated, shader-heavy game), unless they introduce new effects.
It is inconceivable that any game loses anything by supporting Mantle, except development time and therefor, money. That's why Mantle is not suitable for all games; time and money, not technical or genre limitations.
Ok. By saying that it couldn't work, I didn't communicate my point very well. It can work with RTS games, such as Total War. However, I don't feel devs like Creative Assembly would be drawn to Mantle.
It isn't about the graphics performance of a game such as that. Where many RTS games struggle is with the AI calculations. I really enjoyed the latest Total War, but I am sure I do not need to tell you that it is a broken game. Graphically it was quite impressive with a 780 at 1440p. But the AI was broken.
The multitude of coding and all the various layers that goes into that part of the game cannot be done much more efficiently. That's the complex side of the game which sets itself apart from something like an FPS game. And that will probably negate some of the properties of Mantle, even if it could be partially implemented for a boost in graphics performance- the CPU is already tied down.
Is Mantle is going to catch on? At this time i am going to say no. As it's a vendor pacific API and most devs are not going to want to make a direct-x version and a Mantel one. I would rather see developer's focus on open-gl as it's something that works with older hardware from both vendors and is used in Linux.