AMD's Mantle API

Started by
145 comments, last by TheChubu 10 years ago


It's meant to force Microsoft's hand and get them to fix the CPU overhead problems with DX on PC.

That may be the case, but there's no fixing Direct3D without gutting it. They've tried, they've gone as far as they can, and its still not good enough. If we see an API called Direct3D 12 it will be a very different beast methinks.

But that would be fine. Mantle isn't about control or vendor lock-in, its about evolving graphics programming and moving the state of the art forward. Whatever effort AMD will have invested in mantle will serve as prior art to the new status quo, and everyone will be happier for it. Supporting mantle won't be wasted effort now, because it'll just put you in a good position to support the new vendor-agnostic hotness later. Its not even that Microsoft (or the OpenGL ARB) will be more capable of developing a new 3D API, its simply because they are effectively Switzerland, acting as a mediator between politically-tense countries who can't be seen collaborating directly.

throw table_exception("(? ???)? ? ???");

Advertisement
Indeed; any investment with Mantle now will both set you up well AND give a better experience to a group of players and thus potentially more sales (for both you and AMD).

There is very little downside in supporting it which is why DICE are behind it (which means any FB3 powered game) and Thief should be shipping with it - which is an Unreal Engine game so assuming that gets pushed back into UE that's two major engines shipping with support and that alone will force CryTek's hand and Ubisoft too.

Graphics programming is an arms race after all...

I'm very interested in Mantle, my main concern is not performance (still good to have), but doing some experimenting with new render tecniques, months ago I heard rumors about a new forward rendering tecnique from ATI, but found nothing about that, maybe now I'll see if that is possible. Does anyone know if it is possible to get symbols from Mantle even without the SDK? I'd like to try something already.

Peace and love, now I understand really what it means! Guardian Angels exist! Thanks!

I'm very interested in Mantle, my main concern is not performance (still good to have), but doing some experimenting with new render tecniques, months ago I heard rumors about a new forward rendering tecnique from ATI, but found nothing about that, maybe now I'll see if that is possible.

Pretty sure you can download the Forward+ sample from AMD's developer website as set if their Radeon SDK, or whatever it's called.

I'm very interested in Mantle, my main concern is not performance (still good to have), but doing some experimenting with new render tecniques, months ago I heard rumors about a new forward rendering tecnique from ATI, but found nothing about that, maybe now I'll see if that is possible.

Pretty sure you can download the Forward+ sample from AMD's developer website as set if their Radeon SDK, or whatever it's called.

Link

According to this blog entry Microsoft are working on "bringing lightweight runtime and tooling capabilities of the Xbox One Direct3D implementation to Windows":

http://blogs.windows.com/windows/b/appbuilder/archive/2013/10/14/raising-the-bar-with-direct3d.aspx?Redirected=true

I assume that would be fixing the draw call overhead, similiar to what Mantle does.

According to this blog entry Microsoft are working on "bringing lightweight runtime and tooling capabilities of the Xbox One Direct3D implementation to Windows":

http://blogs.windows.com/windows/b/appbuilder/archive/2013/10/14/raising-the-bar-with-direct3d.aspx?Redirected=true

I assume that would be fixing the draw call overhead, similiar to what Mantle does.

It certainly sounds like there is work being done to bring DX.x to Windows (or parts thereof).

My greatest hope for Mantle is that it will pressure the OpenGL ARB into developing an OpenGL 5.0 that doesn't fall massively short, as each major version update has.

I think this is a step in the wrong direction. Portability and hardware abstraction are good things, even if they cost a little performance. You wouldn't want to write 3 or 4 different code paths for 3 or 4 different low-level APIs, possibly with extra hardware-dependent paths, would you

depends on the quality of this 'abstraction'.. If this abstraction is ten times bigger (for example has 10 times more api calls) to use than underlying api its abstraction is not such good. Imo opengl is overbloated,

I would vote for both low level specyfic vendor api and middle layer on top of many vendors equipment

In my opinion there is room for both more lower level APIs like Mantle which one would only use for performance reasons and room for more higher level APIs which one uses if one wants to have shorter production times. It all depends on the application if it makes sense to choose Mantle.

"What? It disintegrated. By definition, it cannot be fixed." - Gru - Dispicable me

"Dude, the world is only limited by your imagination" - Me

This topic is closed to new replies.

Advertisement