Mantle is an alternative to DX as much as a fingernail is an alternative for a toothpick. They both can achieve equivalent results, but their scopes are vastly different.
AMD Mantle API [updating]
- Thread starter MarkoIt
- Start date
Mantle is an alternative to DX as much as a fingernail is an alternative for a toothpick. They both can achieve equivalent results, but their scopes are vastly different.
exactly what i think.
here do not agree: it would be easier porting from a (superset of) dx11.2 to dx11, or from mantle to the dx (because in pc market obviously there are nvidia and intel gpus too, no developer could ignore this of course)? : D
In discussing the future of Direct3D, Microsoft specifically mentions their general commitments to improving Direct3D alongside “bringing the lightweight runtime and tooling capabilities of the Xbox One Direct3D implementation to Windows,”
http://www.anandtech.com/show/7420/...antle-support-names-xbox-one-api-direct3d-11x
http://www.anandtech.com/show/7420/...antle-support-names-xbox-one-api-direct3d-11x
Last edited by a moderator:
From other news DX 11.+ in XB1 is more Mantle than DX 11.2 in PC.
At least 8 millions for DICE it seems
The whole thing is from the "Windows App Builder Blog". And sure, if you are going to write apps that will run on the XB1, it will be on D3D11.2. That's not the same thing as saying that the actual games (ie those things running on the 6 cores and most of the memory) will use generic directx.
And of course Mantle is not on xb1 currently, it's not even done yet - if it makes sense to deliver it as middleware on top of the low level APIs at some point we don't know yet (but probably not really relevant for the AAA development - it's more important to maintain the programming model than the actual calls).
And of course Mantle is not on xb1 currently, it's not even done yet - if it makes sense to deliver it as middleware on top of the low level APIs at some point we don't know yet (but probably not really relevant for the AAA development - it's more important to maintain the programming model than the actual calls).
This makes no sense. Yeah obviously games will still have D3D/OGL. The point of mantle is that it's close enough to consoles' rendering path (doesn't have to be a 1:1 copy) that it allows for easy/painless porting. It is not meant to be a replacement for D3D/OGL but as a supplement.
Last edited by a moderator:
It's a bit opaque at this point. The problem is that most people don't actually know how console APIs differ from DirectX on the PC. I imagine that the order and granularity at which things are done is sufficiently different that Mantle makes porting to PC easier even if it's not the exact same API.
I don't see any reason why Microsoft and the IHV's couldn't provide more direct access to graphics hardware via DirectX. It will be never as close as a custom API targeting a single architecture but they can probably get very close. Modern GPU architectures aren't that different after all.
talking about development costs has a lot of sense: unless you can develop simultaneously in dx and mantle...why spend money to make another version of game running in Mantle, when you can develop in dx for Windows and xbox1 (PS4 is another story)?
The real problem is outside of repi no one knows how different mantle is from consoles/PC.
I just assume AMD has tried to make it as similar as possible. Who knows!Sure but until then...
There are also a few issues with this approach (for the short term at least). Which Windows versions will support this "new D3D"? Will it only apply to "D3D 12" cards or will it support 11.0+/11.1+ cards (more fragmentation...)? Doesn't this defeat the purpose of WDDM? If not, how will this interact with WDDM and still achieve its goal? I'm not confident MS can introduce an elegant solution in the near term. Therefore, I see Mantle at least as a credible short term "solution". And no, blindless extensions in OpenGL don't count (wake me when a AAA game uses it).
But this brings us to the reason why I'm most excited about Mantle. As I eluded to before, I don't care if it fails as an api. It has already achieved its primary goal by forcing MS to bridge the gap (more aggressively) between console apis and pc apis. I'm of the opinion that D3D has stagnated for the past four years (PRT is nice, but not earth shattering...). There's a reason MS went out of its way to claim there's no Mantle support on X1 . This demonstrates to me that they are "feeling the heat". I am glad to see some renewed competition (sorry open source fans, ogl just copies d3d). So while Alex loves to hate on Mantle
razz, I'm glad AMD did it. Even nv and intel fans will receive a long term benefit imo.Same reason why many games support both D3D 9 and D3D 10/11...
Priyadarshi
Newcomer
On Xbox you can directly change the GPU state by setting register values and writing commands into command-buffer. There is a "dx9gpu.h" which contains low-level detail of its Xenos GPU. The presentation : "Stripped Down Direct3D: Xbox 360 Command Buffer and Resource Management" explains some core-to-metal programming stuff on consoles which is different than the standard DX on PCs.
I think Mantle is something very similar, only that now AMD has taken the responsibility of creating such an API (through DX) and lifting the burden off Microsoft.
I think Mantle is something very similar, only that now AMD has taken the responsibility of creating such an API (through DX) and lifting the burden off Microsoft.
Yes, and no. Consoles are a specific case in and of itself. You won't program a PC like that.
Agreed, that's why I added the qualifier "as possible". The idea is to make porting as seamlessly as possible.
And note, we're not talking about "the PC" here "the PC" port is still DX or OpenGL, we're talking about specific GCN hardware targets that shares many commonalities from an architecture and performance optimization point of view. And this is also a two way thing, not a "console to PC port" discussion; many devs will start on the PC and then transition to console, with Mantle you have a better prediction of the behavior of the app on other platforms.
totally agree.
Oh come on, no one needs another api (aka industry standard), everybody needs good ones
if amd doesn't pay for mantle version, there isn't mantle version of a game (that running on the minority part of market)
Then they should stop talking about it and just get on and do it. I'm fed up of Microsofts empty promises to support PC gaming which almost always go no-where.
Pretty much DEAD ON ARRIVAL and irrelevant.
I'm not sure why Mantle not being on either console has any bareing on it's success on the PC. It's still a low level API that will allow developers to extract the maximum performance from PC hardware - just what they've been asking for.
Besides, according to Extremetech:
http://www.extremetech.com/gaming/168671-xbox-one-will-not-support-amds-mantle-and-ps4-is-also-unlikely-is-mantle-doa
Extremetech said:
I completely agree with willard that Mantle is a great thing regardless of whether it's successful or not because it's rattling the right cages.
Also Dave makes an excellent point about developers starting on PC with GCN and Mantle then porting to consoles and DX from there. That makes a huge amount of sense to me. If devs want to start development on the PC as they often do, then using a console style archictecture with a console style API makes far more sense than using different hardware (NV) and a very dissimilar API (DX). Sure you've still got to port back to DX afterwards (or perhaps in parallel) but the advantages are still great.