AMD Mantle API [updating]

[DieH@rd]

Engine programmer of pCARS has commented about Mantle on their forum, he's mostly worried about implementation costs.

However, there isn't any doubt in my mind that Mantle will solve the poor implementation of multi-threaded rendering in the PC space, more so for AMD since they don't even support Driver Command Lists in DirectX! That alone will get them to 70-80% of next-gen console performance in terms of submitting draw calls to the hardware. When DICE reveal their Mantle version in mid November, you'll see scalability that's unheard of in the PC space e.g. some 6 or 8 core demo with submission performance many times that of DX, with lots of pretty charts.

For now pCARS does not have announced Mantle support, however because game will most surely drop PS3/X360 versions and focus on nextgen ports [negotiations between devs and backers is ongoing], eventual Mantle implementation is likely to happen.

 

[lanek]

Johan is completely stoked about Mantle and it could barely be more obvious. Of course if you have a bunch of Nvidia centric guys there like Carmack, at an Nvidia event asking loaded questions then it's difficult not to respond with an appeasing answer. More telling was him making the point about "pitching the idea to any vendor for a number of years. Nvidia was given every chance.

I have sadly to aggree, when i read some comment of Carmack and co during the presentation the Nvidia presentation, im really sad, some of there remark are really a joke, it could have been done by JHH on a interview.. Ok they was here for sell Nvidia stuff, but well...

If i was a game developper, i could maybe not be interested right now in Mantle for my projects, but just because i m a game developpers i will be interested by know more about it, play with it.. Look like someone have take something from him.

 

[MfA]

I posted it multiple times, it always ended at the bottom of a page but the chief engineer behind the XB1 stated in the interview he gave to the DF that they (MSFT) wants to make PC/direct x closer to consoles.

If they announce work on DirectX 12 I'll believe it ... in the mean time every time I hear about bringing the platforms closer together I think they are talking about using the PC for social media games associated with the XBOX games.

I expect NVIDIA will just start heavily pushing for OpenGL support in games.

 

[trinibwoy]

I expect NVIDIA will just start heavily pushing for OpenGL support in games.

You would think nVidia is just as invested in minimizing DirectX overhead as DICE and AMD. I wonder how much actual lobbying is going on behind the scenes and how much Microsoft is really on board.

It sounds like Mantle was simply born out of frustration with the status quo. Kudos to AMD for shaking things up.

 

[firstminion]

I thought Carmack would be more enthusiastic over AMD's SW efforts, they are pushing a lot of things he wants, from textures to unified memory.

 

[Davros]

We have
Mantle, TruAudio, Physx, Cuda, Lightboost, G-Sync, NVapi All of them proprietary.
I'm starting to think this is a bit of a mess

he's mostly worried about implementation costs.

Nvidia said to devs "we will port your game physics to Physx" maybe amd need to do the same

more so for AMD since they don't even support Driver Command Lists in DirectX! That alone will get them to 70-80% of next-gen console performance in terms of submitting draw calls to the hardware.

Do Nvidia support Driver Command Lists in DirectX ? Do they have 70-80% of next-gen console performance in terms of submitting draw calls to the hardware ?

 

[xDxD]

here the complete video about this great developers and their opinions on mantle: http://www.youtube.com/watch?feature=player_embedded&v=3exPJu_F8xk

 

[liolio]

If they announce work on DirectX 12 I'll believe it ... in the mean time every time I hear about bringing the platforms closer together I think they are talking about using the PC for social media games associated with the XBOX games.

I expect NVIDIA will just start heavily pushing for OpenGL support in games.

Well I agree with you talking is free. Tough those guys were pretty explicit:

We also took the opportunity to go and highly customise the command processor on the GPU. Again concentrating on CPU performance... The command processor block's interface is a very key component in making the CPU overhead of graphics quite efficient. We know the AMD architecture pretty well - we had AMD graphics on the Xbox 360 and there were a number of features we used there. We had features like pre-compiled command buffers where developers would go and pre-build a lot of their states at the object level where they would [simply] say, "run this". We implemented it on Xbox 360 and had a whole lot of ideas on how to make that more efficient [and with] a cleaner API, so we took that opportunity with Xbox One and with our customised command processor we've created extensions on top of D3D which fit very nicely into the D3D model and this is something that we'd like to integrate back into mainline 3D on the PC too - this small, very low-level, very efficient object-orientated submission of your draw [and state] commands.

WE can only hope but I'm neither a lobbyist nor an activist of some form but some people on the web are good at it, when Mantle launches and if it delivers significant improvements I'm sure some people could manage to make a lot of noise, a noise MSFT may have to take in account.
The same people could also attack AMD (the stupid move) instead, the latter is more likely, that is a sad matter of fact.

 

[Davros]

that video
http://www.youtube.com/watch?feature...&v=3exPJu_F8xk
what does he say about 9:45 "the vision for mantle it can become ???????? api also

 

[pjbliverpool]

That conversation certainly puts the possibility of an nvidia Mantle equivalent in great doubt.

 

[flopper]

I thought Carmack would be more enthusiastic over AMD's SW efforts, they are pushing a lot of things he wants, from textures to unified memory.

The way its meant to be payed from Nvidia.
Make no mistake those guys are so full of it.

 

[DieH@rd]

that video
http://www.youtube.com/watch?feature...&v=3exPJu_F8xk
what does he say about 9:45 "the vision for mantle it can become ???????? api also

I watched different live stream, he said "open" API there.

 

[Davros]

ok thanks

 

[jimbo75]

Not much chance of that happening but it's the diplomatic answer. :smile:

 

[Dominik D]

The fact remains that AMD paid DICE a large amount of money for Battlefield 4, and that includes Mantle support. According to AMD at this very forum, it will take more development work to optimally code for Mantle compared to DirectX or OpenGL. More development work does not come for free.

This is false reasoning. Mantle is an expense that gives you something extra, not development work (cost) that's required to be on par with something else. DICE is working on one of the 3-4 most advanced engines out there and it's their prerogative to stay on top of the cutting edge technology. It so happens that they work with AMD to deliver something spectacular - so be it. Assuming DICE/EA was paid to use Mantle because they have business relationship WRT to cross-promotion of products is bull.

I used to work as a journalist in my previous life and statements like these enrage me. No matter how well I was trying to do my work, the moment there was an intersection of what I wrote about and who paid for adds, people were crying paid content. Sometimes I wish that was the case - at least I would be reimbursed for having to deal with people who spew BS like this.

Also stating something like this and calling it "fact" should be penalized with hedgehog in your pants.

 

[Arwin]

What I took from this video is:

- Mantle allows for testing new ideas, and finding bottlenecks in rendering on PC in a way that was not possible before, and that will help inform all other platforms as to where they should spend more effort to relieve the biggest bottlenecks
- Ditto for testing new designs and rendering pipelines, for perhaps a new more efficient higher-level API that is based on the current situation (DirectX and OpenGL have been "around"), where GPU and CPU converge and both are (sometimes very) many core systems. There is a chance that Mantle itself will evolve in this direction and become 'Open'.
- Mantle allows multi-platform developers to leverage knowhow and optimisations from the console space onto PC. This is a direct consequence of AMD having won both console contracts.
- Despite his business sense saying 'this is not a good idea for us', I couldn't help but shake the feeling that Carmack was convinced that there could be very many benefits to using Mantle (they made a pretty big list just by going back and forth over current 'problems' for a minute or so)

Biggest counter argument against Mantle I heard so far is that Mantle is still too different from PS4/Xbox One APIs to make leveraging console code on PC a 'quick' win, and therefore a good business decision. So this is where AMD might need to do an 'Nvidia' and strategically volunteer to do the work for some key players.

 

[flopper]

More interested if the Mantle api makes amd cpu´s be a viable option for BF4 and other games with the Mantle api support.
if it changes the cpu/gpu balance an 8core 8350 might become the best choice for BF4/Mantle.

 

[Arty]

I truly feel sorry for Johan Andersson. Nvidia flew him out there, treated him like a celebrity and then landed that low-blow. It was all pre-planned, knowing Rein was swimming well before this event in Nvidia's deep pockets. I dont think it was all a co-incidence.

 

[ams]

This is false reasoning. Mantle is an expense that gives you something extra, not development work (cost) that's required to be on par with something else.

That is a ridiculous assertion. Mantle is an IHV-specific API that requires more development work to achieve optimal results compared to DirectX and OpenGL, period. The main benefits of Mantle will only be realized in GCN-equipped systems, nothing else. In this day and age, developers need to maximize graphics performance and stability across a wide range of platforms including: PC, Xbox360, PS3, 3dS, PSP, iOS, Android, Xbox1, PS4, etc. On the PC side, the majority of users with integrated or discrete graphics use Intel CPU/GPU or Intel CPU + NVIDIA GPU. Xbox360, PS3, 3dS, PSP use non-GCN GPU architectures (Xbox360 is the only system with AMD inside). iOS and Android use non-GCN GPU architectures (none with AMD inside). Xbox1 and PS4 are GCN-equipped (with AMD inside), but are customized for the respective closed platforms, and use their own low level and high level graphics API rather than Mantle. So to suggest that game developers should incorporate Mantle and spend extra time optimizing for Mantle for free on top of what they are already doing for DirectX and OpenGL is crazy. Keep in mind that Johan @ DICE hopes that Mantle will push forward the state of DirectX/OpenGL/OpenGL ES API's, and even he does not want any IHV-specific graphics API to propagate over time.

 

[Dominik D]

Mantle is an IHV-specific API that requires more development work to achieve optimal results compared to DirectX and OpenGL

Optimal and better, not equal.

The main benefits of Mantle will only be realized in GCN-equipped systems, nothing else.

Which is cross-promoting Battlefield. Seems like investment in target market.

On the PC side, the majority of users with integrated or discrete graphics use Intel CPU/GPU or Intel CPU + NVIDIA GPU.

Majority of PC owners will get gimped experience anyway with their low-to-mid-tier GPUs. So you could ask: why invest in DX11.2 features? Or vendor-specific OGL extensions? Answer is in my previous post: DICE is developing cutting edge technology and BF is a showcase of their efforts. In their case investment in Mantle does make sense.

So to suggest that game developers should incorporate Mantle and spend extra time optimizing for Mantle for free on top of what they are already doing for DirectX and OpenGL is crazy.

Equalizing gaming landscape, putting all games in one bag and using generalizations to draw bogus conclusions about one specific game (or one engine) is much, much crazier.

Keep in mind that Johan @ DICE hopes that Mantle will push forward the state of DirectX/OpenGL/OpenGL ES API's, and even he does not want any IHV-specific graphics API to propagate over time.

That's not exactly what he said. This is your interpretation of his "I don't want every IHV with its own standard" statement. You've dropped the context of what he said to support your claim which is not OK.