Another Game Studio Backs AMD's Mantle API

[phoronix]

Phoronix: Another Game Studio Backs AMD's Mantle API

There's another game studio now backing AMD's Mantle graphics rendering API that aims to be faster and easier to implement for games than OpenGL. However, we're still waiting for AMD Mantle on Linux...

http://www.phoronix.com/vr.php?view=MTUyNTc

[godlike_panos]

Faster probably, easier I don't think so. Mantle is quite low level and that requires to do more things yourself (eg dependency tracking). Overall it sounds like a nice API because it gives more freedom and predictability to developers but I don't think other vendors will adopt it in it's current form.

[TemplarGR]

It is nice to see Mantle getting more support, but Rebellion is not actually known for its quality games...

[who_me]

The company responsible for some of this year's most terrible games is excited about Mantle. Ohhh, I'm soooooo excited too! NOT.

[entropy]

Granted, considering the poor state of the Radeon R9 290 on Linux, AMD might not be in an immediate rush to push Mantle
on Linux but it's great to see game studios taking a greater interest in the API spawned by AMD.

I don't get this.

*If* we assume Mantle is already in a good shape and only needs some final polish,
wouldn't that be a great chance for AMD to circumvent the OpenGL misery (see Metro: LL, etc.)?
And as it is a low-level API, it sounds like it doesn't require much work to bring it to Linux.(?)

[BlackStar]

Faster probably, easier I don't think so. Mantle is quite low level and that requires to do more things yourself (eg dependency tracking). Overall it sounds like a nice API because it gives more freedom and predictability to developers but I don't think other vendors will adopt it in it's current form.

It cannot be harder than OpenGL, that's a given. Not only do you have to contend with API madness, you have to fight against drivers all the way through (does any driver actually implement UBOs to spec yet?)

So yeah, Mantle could be a breath of fresh air. Although considering it's AMD, it will probably be half-assed and half-supported in the end.

(How about this: release it as an open specification and make sure other people can implement it. Keeping it behind closed doors until 2015 is not going to help anyone.)

Edit: have they said anything about Linux support?

[mark45]

I don't get this.

*If* we assume Mantle is already in a good shape and only needs some final polish,
wouldn't that be a great chance for AMD to circumvent the OpenGL misery (see Metro: LL, etc.)?
And as it is a low-level API, it sounds like it doesn't require much work to bring it to Linux.(?)

Good questions, but the only answer for now: time will tell, no one knows for sure.

[Ferdinand]

Faster probably, easier I don't think so. Mantle is quite low level and that requires to do more things yourself (eg dependency tracking). Overall it sounds like a nice API because it gives more freedom and predictability to developers but I don't think other vendors will adopt it in it's current form.

As far as I understand only engine developers need to target Mantle. Game developers just use the engine and the tools that come with it. There are only 4 really popular engines so seems pretty easy to me.

[johnc]

Edit: have they said anything about Linux support?

DICE is "hopeful" that Mantle will land on non-Windows systems some day, but there's been no indication that any effort is occurring to make that happen.

AMD has been 100% mum on the topic of course.

Mantle will certainly give Windows a performance advantage over, say, SteamOS.

[mrugiero]

Faster probably, easier I don't think so. Mantle is quite low level and that requires to do more things yourself (eg dependency tracking). Overall it sounds like a nice API because it gives more freedom and predictability to developers but I don't think other vendors will adopt it in it's current form.

I think the "easier to implement" statement is actually referred to drivers, as you implement more basic operations, it might be easier to implement.