AMD Mantle API [updating]

[Gipsel]

This is a classical "post hoc" logical fallacy. Xbox One has been in active development for the last 2.5 years, long before the Mantle announcement.

I thought you just linked to a description of a logical fallacy?
The MS statement came after being questioned about Mantle (and its relation to the XB1). That establishes a causal link. Without that, they wouldn't have made that statement. It stands to reason, that they wouldn't have said that they are checking what aspects of the XB1's DX11.x could be used for that purpose. And I'm quite sure that MS also takes a careful look at Mantle and its concepts to evaluate their future plans with DX. Anything else would be foolish.

 

[DmitryKo]

I still wouldn't believe the claim that a magically faster runtime on Xbone makes the same code go way faster without proof.

This claim is made by Microsoft developers who actually design the Direct3D runtime and tools for the Xbox One.

The MS statement came after being questioned about Mantle (and its relation to the XB1). Without that, they wouldn't have made that statement.

This still doesn't mean the actual decision to port the improved D3D11 runtime was made after the Mantle announcement - unlike the recent decision to base the design of Direct3D 12 around Mantle concepts.

 

[Gipsel]

This still doesn't mean the decision to port the improved D3D11 runtime was made after the Mantle announcement.

But it was said as a direct reaction of being questioned about Mantle. And I claimed nothing more than that the statement was a reaction to Mantle.

Unlike the decision to base the design Direct3D 12 around Mantle concepts.

If you think MS does this, why do you argument against Mantle pushing the development? I'm confused.

 

[DmitryKo]

I claimed nothing more than that the statement was a reaction to Mantle.

Development of a lighweight Direct3D runtime for the Xbox One was not.

If you think MS does this, why do you argument against Mantle pushing the development?

Which development exactly?

One thing is optimisation of the existing Direct3D 11 runtime and API. This is a routine process that happens with each minor release.

Designing a new Direct3D 12 API that should closely follow the concept of Mantle and in the process probably break any compatibility with D3D11 (except maybe for HLSL shaders) is a whole another story. Last time this happened was 10 years ago with Windows Vista and Direct3D 10 and all recent releases of the API have been built on top of Direvct3D 10.

 

[Gipsel]

Development of a lighweight Direct3D runtime for the Xbox One was not.

Of course not. But bringing these ideas to the PC is something different.

Which development exactly?

That MS has some stronger incentive to come up with a Mantle like D3D version or extension.

 

[Malo]

What incentive? They've never really cared much about PC gaming due to Xbox market. They don't want people buying games for the PC when it is available on Xbone.

 

[swaaye]

DirectX has been about keeping people on Windows and preventing OpenGL from becoming the game API of choice (portability danger). There are some new threats to Windows and their API stranglehold now.

 

[DmitryKo]

That MS has some stronger incentive to come up with a Mantle like D3D version or extension

Doh. I am repeating the same two statements a fourth post in a row. What's the problem - is my English not good enough?


"Lightweight runtime" for Direct3D 11 is not a "Mantle-like extension", it has been developed for the last 2.5 years, so it cannot be influenced by Mantle. The decision to introduce these improvements to the PC is quite natural, and even if the announcement of this decision was influenced by Mantle, this doesn't mean the decision itself was, unless someone at Microsoft explicitly confirms it.

However a new version of D3D based on Mantle, if that is true, is undoubtedly being influenced by, well, the Mantle announcement and its warm reception in the developer community.


That said, a curious side effect of both announcements could be an introduction of entirely new display driver model, something like WDDM 2.0, which would help materialize both "lightweigh runtime" for Direct3D 11 and new "Mantle-like" Direct3D 12 in Windows 9.

I suggest we curb further speculation and wait for the actual GDC papers to go online.

 

[gongo]

Good news AMD bros! The new drivers....fixed my cold boot problems _!_ ..so now whenever i turn on my PC in the morning and run a 3D game/app, no half clocking whatsoever..with or without MSI AB..it works proper!

Kinda sad though i am excited over something this basic, but it has plagued me for a month, got so annoyed i played less games and went out to the park more often ..

Oh wait btw...new drivers same problem...if i chose not to install HDMI drivers..CCC will fail to load...so there Dave another annoyance to look into...

 

[willardjuice]

"Lightweight runtime" for Direct3D 11 is not a "Mantle-like extension", it has been developed for the last 2.5 years, so it cannot be influenced by Mantle.

This is the crux of the debate. I don't think this statement is true. I think what's been in development for 2.5 years is a version of d3d11 that has been specifically tailored to gcn (and not a "generic" lightweight version).

I'm not saying they couldn't apply some of what they learned on x1 to d3d's next iteration or that they only started thinking about lower level access after mantle. However, it seems likely to me (based on their announcements and the timing of their announcements) that mantle has at least changed their priorities a bit. This is a good thing. :smile:

 

[itsmydamnation]

maybe XB1 getting its butt kicked by PS4 help change their priorities /runs

 

[MfA]

This is a classical "post hoc" logical fallacy. Xbox One has been in active development for the last 2.5 years, long before the Mantle announcement.

He said he regarded it so, not that it must be so ... certainty can only be had in math, in the real world evidence is never complete.

I for one agree with him, AMD said "there will be no DirectX12 at any time in the future" and Microsoft is silent ... AMD releases Mantle and shit gets moving. Microsoft has been schizophrenic on PC gaming ever since the XBOX started ... hell this is Microsoft, they might not even plan to release anything good on the PC. There is a real possibility this is just an embrace and extinguish campaign relying not on actually providing an alternative but merely promising one to make developers weary of costly investment in either Mantle and/or OpenGL on the PC.

 

[Jawed]

Apparently there are some here who think Mantle has been in development for 6 months

 

[Malo]

Yeah it's been at least 2 years according to release information.

 

[MfA]

Apparently there are some here who think Mantle has been in development for 6 months

Because management decisions are never made on the fly based on what's in the press at any given time ...

 

[silent_guy]

Because management decisions are never made on the fly based on what's in the press at any given time ...

What's the relevancy of this statement? Of course management will often react to stuff in the press. They'd be stupid not to. That doesn't mean that the results of those decisions are announced way before they are ready.

Do you think that Mantle took less than 6 months to come up with, evangelize, and develop?

 

[MfA]

No, but I think Microsoft's new zeal for DirectX development could have easily been started in that time.

 

[DmitryKo]

what's been in development for 2.5 years is a version of d3d11 that has been specifically tailored to gcn (and not a "generic" lightweight version).

The point is, a GCN-tailored runtime would also benefit Fermi/Keple, given their overall architectural similarities. But that should probably require a new driver model, which reduces the driver overhead by streamlining the DXGI API and data formats.

it seems likely to me (based on their announcements and the timing of their announcements) that mantle has at least changed their priorities

Such announcements are typically made at a late stage of actual development, when the feature in question at least partially implemented and tested.

AMD said "there will be no DirectX12 at any time in the future" and Microsoft is silent ... AMD releases Mantle and shit gets moving

Of course if they say "Direct3D12 will be a new Mantle-like API", that's a confirmation in itself.

 

[Blazkowicz]

I had written something about OpenGL but on the "Why was glide faster...?" thread, so I'll quote myself :

All the new features, core and extensions in GL 4.4 are named GL_ARB_something and three of these new things are advertised as reducing the CPU overhead.
https://www.opengl.org/documentation/current_version/

For now AMD and Intel only support up to OpenGL 4.3. I tried to google things, and could find this

According to Graham Sellers, OpenGL guy at AMD, the red team will be supporting this open API with some high performance extensions that will offer almost similar performance to AMD’s upcoming API, Mantle.

http://www.dsogaming.com/news/amd-aims-to-give-opengl-a-big-boost-api-wont-be-the-bottleneck/

It sort of implies AMD is working on OpenGL 4.4 support but no version is named explicitly. OpenGL extensions are evoked, but without explicitly saying whose extensions are they.

and one of Graham Sellers twitter post says "I'm saying that for a modern OpenGL application, API won't be the bottleneck. You'll hit HW limits first."

So in a way AMD said OpenGL will be a Mantle-like API .
Thoughts?

/edit : Graham Sellers did not directly say that, and this stuff was mentioned briefly on page 8 of this thread.
Also nvidia has been supporting OpenGL 4.4, but that didn't make big headlines. Maybe when SteamOS and Tegra K1 are released we'll hear again.

 

[Zeross]

Modern OpenGL doesn't really reduce draw call overhead per-se, however these extensions allow you to draw more things inside the same batch. The end result is the same but I think the distinction in how this is achieved is important.

Moreover low overhead is not the only strong point of Mantle. Since this is a cleansheet design it allows to present the pipeline the way it is now and to have a better control on the memory like on consoles. OpenGL (or D3D for the matter) abstract all of that through layers of abstraction.

Don't get me wrong: I think that OpenGL is on the right rack and it is amazing to see what was achieved during the last five years. When the Longs Peak redesign went nowhere I admit that I lost all hope in Khronos but they proved me wrong.