It seems Mantle can be easily ported to DX12

[CAT-THE-FIFTH]

Supposedly:

http://wccftech.com/mantle-api-porte...d-white-paper/

Good news if true.

[kompukare]

The next question is of course if Mantle is so similar to what DX12 will be, who if anyone is copying who? And are any of these API architectures covered by patents?

Because Mantle for Linux (especially SteamOS) would be an obvious idea but for some reason AMD have not said anything. So unsure if that's because AMD don't have the manpower, AMD don't want to offend Microsoft, or that they are afraid that some of the basis of how Mantle works could be covered by some Microsoft DX patent.

Certainly if SteamOS are going to be dependent on the full OpenGL it could be rather messy. Sometimes breaking backward compatibility and starting from a (relatively) clean sheet is a good idea but it seems the way OpenGL and the Khronos group work is the opposite of that. Guess there is OpenGL ES.

[kalniel]

Low barriers are always good news, but there's only a small window where it'll be advantageous to code for mantle now - depending on MS's plans for DX12. If MS release it for Win 7 then devs will probably jump straight to it. If they release it for Win 9 only then Mantle becomes way more useful.

[CAT-THE-FIFTH]

Low barriers are always good news, but there's only a small window where it'll be advantageous to code for mantle now - depending on MS's plans for DX12. If MS release it for Win 7 then devs will probably jump straight to it. If they release it for Win 9 only then Mantle becomes way more useful.

OTH, if porting over to DX12 is not a massive effort as purported by AMD,then its more likely that devs will likely to implement Mantle as part of the AMD gaming partnerships and console wins,and its more likely more devs will implement it now rather than wait for DX12 first:

http://www.gamersnexus.net/news/1432...new-developers

It could explain why dozens of devs have signed up for Mantle now.

IMHO,the whole point of Mantle was to help AMD CPUs look more competitive,until they get their next design out in 2016,and by the time DX12 is out loads of engines will have Mantle support I suspect.

[scaryjim]

If mantle *is* - as a lot of us have speculated - based on the low-level API for current gen consoles, it wouldn't be a big surprise that DX12 is very similar. IIRC there's been caselaw established that you can't protect an API, only an implementation of it, so there shouldn't be any issues in that regard. And given MSes - as I understand it - fairly open approach towards mono, I can't see them being too hostile towards API porting. After all, the more games that are DX12 ready and optimised when it is released - particularly if it is a Windows 9 exclusive - the better for MSes marketing department...

EDIT:
Just a quick though on open mantle: Mantle being open is only useful if intel & nvidia implement it, so its success is going to depend entirely on uptake from competing companies. OTOH if AMD get Mantle working on Linux, sudenly they've given devs a great way to target linux and windows really easily, and with an easy port to DX12 if that's the way they want to go in the future. That might just be enough to force nvidia and intel to adopt mantle on Linux, which would be a huge win for AMD (assuming they can get their Linux drivers sorted out, that is ).

[Agent]

Thing is "effort" and "low / medium" on a slide from the company means very little, it's just another form of marketing.

What's really needed is a man hour calculation against different project sizes.

The difficulty rarely comes from just using the API. It's implementing it within an existing codebase (testing, bugs, deployment and so on...). If AMD had said "It look Epic x man years to implement Mantle into UE4" or "<big title> ported to Mantle in x man years" then it becomes more interesting.