AMD Mantle API [updating]

Because DX12 won't be here for nearly 2 years?

 

Plus, given the apparent similarity of Mantle and DX12, much of the time invested into Mantle should be beneficial to future DX12 work. Otherwise Crytek wouldn't have announced Mantle support on the same day that MS unveiled DX12.

 

That's a very good point and may well breath some life into Mantle over the next 2 years if developers see it as a kind of test bed for DX12.

 

But why they have to waste time and money on something that has a deadline?
I think it's best to use the resources in the OGL or DX12.

 

Who told you Mantle has a deadline?

 

DX12 would probably be only supported on 8.1 or later. With Mantle MSFT cannot force you to buy a new OS to use new graphics features.

I think he wanted to say that Mantle is going to be dead after DX12 is released.

 

That isn't a usable metric for this situation.
Most mobile SoCs may eventually bring D3D12 compliance but they sure as hell won't use it. Unless you think Android/iOS will adopt DirectX

Spoiler

(won't happen)

, Windows Phone finds a way to drastically increase its 3% marketshare

Spoiler

(highly unlikely)

or Windows RT tablets suddenly become a huge success

Spoiler

(LOL)

, mobile SoCs shouldn't even count for the equation.


The best metric should be amount of gaming devices or at least devices being used for gaming.
And -->If<-- AMD convinces developers to use Mantle for both consoles and some major PC titles, then we'll have all xbones + all ps4s + all gaming PCs with GCN GPUs versus all gaming PCs with nVidia and Intel GPUs.


This could put all TWIMTBP/Gameworks-infested dev teams (and Unreal Engine 4 itself) in a very uncomfortable situation, or having to make some very difficult decisions.
Say you're a dev team leader planning to make a multiplatform game.
Will you use nVidia's free engineers, giveaway hardware and (reportedly) better framework for the PC version and then add costly and time-consuming ports for the new-gen consoles? Or will you just use Mantle for everything, getting fast and cheap ports for all platforms, and then rush out a DX11 version for the gaming PCs without a GCN GPU?

Or maybe it's DX12 on PC who will be DOA if AMD, Valve, Apple and Google get their way.

 

This seems monumentally unrealistic to the point of not even worth considering.

XB1 will already have it's custom version of DX12 which will be at least as efficient as Mantle for their architecture as well as being fully supported bu Microsoft themselves. Would Microsoft even allow the use of Mantle on XB1?? While Sony have got their own API's and have never made so much as a hint that they are considering supporting Mantle.

On top of that, Mantle's a PC API and there's never been any mention of it even being compatible with the consoles (without modification) as far as I'm aware.

Once DX12 releases it will basically be all Windows 9 (and possibly 8) gaming PC's + XB1 vs all AMD gaming PC's on older windows versions. PS4 will stand on it's own. It will make more sense at that point (assuming they bother coding two API paths at all) for developers to have a fast DX12 path for all PC gamers on Windows 9 (and possibly 8)and a slow DX11 path for all PC gamers on older Windows. That way all gamers regardless of hardware are treated equally. If they forego DX12 in favor of Mantle then all Nvidia owners (more than 50% of the discrete market) are locked out of your performance improvements.

 

Support beyond Windows is something DICE might want, as it might want more than non-committment from AMD.

If you want to get conspiratorial, Microsoft's decision to hew so closely to Mantle's concepts might have come with a little agreement about how hard AMD is going to push non-Windows Mantle.

 

If DX12 is really close of Mantle as some suggest it, all the work done on it, will benefit to a certain extend their developpement for DX12. I can even imagine the optimisation, the work and time passed then on DX12 optimisation, developpement, drivers etc will benefit the other way around Mantle.

Speaking about deadline now, when we cant compare both yet is a bit early. Till 2015, we have too no idea what direction will take Mantle.

 

The big picture here is that Mantle is not really that earth shattering as of yet. It's application so far brings marginal performance improvements at best(At the high end). With heavy dependance on drivers, some image quality quirks, high memory consumption, and cases of unstable frames, and still it works on some cards and not others.. And unless all of this is ironed out soon, it will be hard to make a case for Mantle against DX12 even during the time before DX12 release.

 

1 - It makes sense from a developer standpoint
2 - Well, when you don't say no...

And won't be available to any game being released throughout the next 2 years.

Maybe the question should be: can Microsoft or Sony forbid using Mantle for their consoles?

 

There's a kernel driver component to Mantle.
As platform owners, they have control over what may exist in the priviledged part of the system, unless we think AMD can play the role of a corporate geohot without massive consequences.

Edit:
Come to think of it, as platform owners of a walled garden, they could also just disallow or refuse to certify software that uses it as well.

 

Technically, they can. But then there's the pressure from major publishers that would love to save money and development time if their dev teams could focus on a single API.
Neither Sony or Microsoft seem to be in a comfortable enough position to arm-wrestle their way into publishers/developers (i.e. like Sony was during the PS2 cycle).

Sony is in dire need of profits and Microsoft has been losing the PR war and the shipped units war by a large margin.
If TakeTwo says "we want to make GTA VI in Mantle", EA says "we want to make Dragon Age, Mass Effect, Battlefield, etc. in Mantle", what will Sony and Microsoft do?

Besides, if Sony and Microsoft really wanted to make sure they'd make a fortress out of their closed gardens, I don't think they would consider using the same CPU and GPU architectures, coincidentally the same architectures that had been used in the PC.

 

That's more of a question of whether they should, not whether they can.

 

Semantics
But yes, my phrasing should have been "Can Sony and Microsoft afford to forbid using Mantle for their consoles".


I think adopting Mantle for consoles is just the natural evolution of what both companies started by adopting identical CPU and GPU hardware.

Development costs for AAA titles have been ballooning way too fast. To me, it makes sense that Microsoft and Sony would mutually agree to ease on the time and costs of portability between both consoles. Make that a neutral ground and then compete with 1st party exclusives, services and peripherals.
That way, less 3rd parties - which both Sony and Microsoft are always in dire need of - will file for bankruptcy and we can keep getting AAA titles without having to turn the whole industry into F2P models.

 

AAA are expensive, but is the graphics API a major compared to everything else? For a well designed engine, you'd expect this to be encapsulated.

 

That assumes that Khronos will not match DX12. They are legendery for their inaction in the DX10 era, but have cleaned up their act quite a bit since then. Although I won't be holding my breath for them to do this.

 

Huh? AMD has no technical way to make Mantle available to console developers if Microsoft do not embrace it.

AMD has no control over either consoles beyond licensing the silicon and providing some engineering and software support. Microsoft control the firmware and the development environment, and they will implement only those APIs as they see fit.

AMD cannot sideway Microsoft and implement Mantle on their own, by making it a display driver extension like they did in Windows. There are no "drivers" on consoles, i.e. DLLs which export some OS-defined device driver API. There is statically linked code which contain the current version of the OS loader and a console version of the Direct3D 11 runtime which is statically linked to low-level code that talks to the actual GCN hardware.

Inaction? OpenGL lost all relevance on the PC gaming market since 2007, when Kronos Group dropped the long-awaited Longs Peak (OpenGL 3.0) proposal, which was exactly what D3D12 is about right now (totally new pipeline and object model with immutable pipeline states), and instead gone the way of maintaining hundreds of optional extensions, just because CAD vendors were not willing to rewrite their rendering engines or maybe some major IHV did not want to rewrite their drivers, whatever. They were unable to address the threats from D3D10/11, and now with Mantle and D3D12 they have lagged behind all chances of recovery. Some extra extensions for high-performance rendering are not going to change the direction the API is going.

 

True epic cant even sell/license console compatible builds of ue4