The AMD Mantle Thread

[OCGuy]

Wrong: PhysX is closed and AMD could not create support even if they wanted too. And it's even worse. NV drivers don't allow run hybrid system in which the NV card is for physx only.

mantle is open. Anyone can create their own implementation of a mantle driver.

No, no it isn't. Anyone can create hardware that has features similar to GCN, but that is a little silly.

Mantle's functionality requires GCN architecture at this point. Don't spread the rumour that that terrible WCCF site or whoever posted stating it works the same on AMD and NV.

For the record I didn't see anything showing AMD pushing that rumor......

 

[Skurge]

No, no it isn't. Anyone can create hardware that has features similar to GCN, but that is a little silly.

Mantle's functionality requires GCN architecture at this point. Don't spread the rumour that that terrible WCCF site or whoever posted stating it works the same on AMD and NV.

For the record I didn't see anything showing AMD pushing that rumor......

You have it wrong. The architecture does not need to be similar. It just needs to have the minimum feature set that Mantle requires. Like DX. Kepler may have the required features or it may not. We don't know. Or nVidia just writes and driver and the devs use which ever features work best on kepler. Theoretically making kepler run better with Mantle than DX. But at this point we aren't sure.

But devs and AMD have said it does not require GCN. That is why whatever the next architecture from AMD is, it will work with Mantle as it will have the same features as GCN plus more. Which Mantle will also make use of.

I say there is a probably a minimum set of features need for Mantle, that is probably why pre GCN cards do not support it. Obviously AMD can't say it will work with kepler cause they didn't build kepler so they wouldn't know if it would work or not. That's what I got from the slides and interviews anyway.

If you had paid read the slides or watched the videos you would know all of this.

 

[blastingcap]

You have it wrong. The architecture does not need to be similar. It just needs to have the minimum feature set that Mantle requires. Like DX. Kepler may have the required features or it may not. We don't know. Or nVidia just writes and driver and the devs use which ever features work best on kepler. Theoretically making kepler run better with Mantle than DX. But at this point we aren't sure.

But devs and AMD have said it does not require GCN. That is why whatever the next architecture from AMD is, it will work with Mantle as it will have the same features as GCN plus more. Which Mantle will also make use of.

I say there is a probably a minimum set of features need for Mantle, that is probably why pre GCN cards do not support it. Obviously AMD can't say it will work with kepler cause they didn't build kepler so they wouldn't know if it would work or not. That's what I got from the slides and interviews anyway.

If you had paid read the slides or watched the videos you would know all of this.

+1

 

[OCGuy]

You have it wrong. The architecture does not need to be similar. It just needs to have the minimum feature set that Mantle requires. Like DX. Kepler may have the required features or it may not. We don't know. Or nVidia just writes and driver and the devs use which ever features work best on kepler. Theoretically making kepler run better with Mantle than DX. But at this point we aren't sure.

But devs and AMD have said it does not require GCN. That is why whatever the next architecture from AMD is, it will work with Mantle as it will have the same features as GCN plus more. Which Mantle will also make use of.

I say there is a probably a minimum set of features need for Mantle, that is probably why pre GCN cards do not support it. Obviously AMD can't say it will work with kepler cause they didn't build kepler so they wouldn't know if it would work or not. That's what I got from the slides and interviews anyway.

If you had paid read the slides or watched the videos you would know all of this.

Providing the data helps the discussion. AMD is saying it "supports" other vendors, but functionality is limited to GCN. If it was just a driver, all AMD cards could use it. The slides say they "hope" vendors in the future will use it, meaning NV/Intel would have to invest money to get it to work, meaning it wont work.


http://vr-zone.com/articles/mantle-shows-promise-proprietary-nature-still-achilles-heel/64273.html

Mantle shows promise, but its proprietary nature is still its achilles heel

Multi-vendor support is a touchy subject when it comes to Mantle. On stage, Andersson said that with Mantle the “pink elephant in the room” was multi-vendor support as Mantle’s functionality is only limited to its own hardware that runs on the GCN architecture. As of now, Mantle is supported on nothing more than AMD’s recent cards that have GCN.

Originally AMD implied that Mantle&#8217;s would serve as a bridge between the PC, PS4 and Xbox One to make ports easy on the common architecture. When Sony and Microsoft announced that they would not support Mantle, AMD denied that this was ever the case (though its friendly press still say otherwise). ( <---- Link to AT article in original)

On a side note, Edit: I may have misread it as an attack. Oh well, no tech-site wars today...maybe next time. the author takes a dig @ AT/Ryan Smith in the following paragraph due to the initial thoughts by basically everyone that Mantle would be in consoles as well as PC. I do not believe that is fair, as everyone was thinking the same thing at the time. Speculation always runs the risk of looking foolish later.....he is man enough to leave the article up. Did I just uncover this or am I late to the party?

 

[smackababy]

On a side note, the author takes a dig @ AT/Ryan Smith in the following paragraph due to the initial thoughts by basically everyone that Mantle would be in consoles as well as PC. I do not believe that is fair, as everyone was thinking the same thing at the time. Speculation always runs the risk of looking foolish later.....he is man enough to leave the article up.

I was one of the few that said Mantle wasn't being used in consoles around the time of the article and quite a few people were responding with "nut uh! Ryan says so".

The principles behind the design for Mantle and for consoles is very similar, but we don't know exactly how similar. Using OO in Java and C++ have similar concepts, but there is a massive difference.


Is there any real news on this besides a bunch of "it's better because x says so" yet? Any real data or tech demo?

 

[Paul98]

I am excited for the SDK to come out, I find this sort of programming interesting.

 

[Skurge]

Providing the data helps the discussion. AMD is saying it "supports" other vendors, but functionality is limited to GCN. If it was just a driver, all AMD cards could use it.


http://vr-zone.com/articles/mantle-shows-promise-proprietary-nature-still-achilles-heel/64273.html

Mantle shows promise, but its proprietary nature is still its achilles heel





On a side note, the author takes a dig @ AT/Ryan Smith in the following paragraph due to the initial thoughts by basically everyone that Mantle would be in consoles as well as PC. I do not believe that is fair, as everyone was thinking the same thing at the time. Speculation always runs the risk of looking foolish later.....he is man enough to leave the article up.

That doesn't refute what I said. Like I said. Go through the slides and you wouldn't need to have other people have to explain it to you.

Mantle is not proprietary. This guy is the 1st person to claim it is and I don't know how much he knows about anything tbh. After using Fudzilla as a source and dissing Ryan.

 

[krumme]

Strawmen and threadcrapping as usual from ocg and smbaby. As if Ryan didnt knew what he was talking about.

Who cares if ms and sony say they support an api for the pc? How is it relevant?

Mantle is getting more support from devs each day. Its getting in to all new engines. Who could have guess that momentum. Can someone actually find a single new engine comming 2014 that wont have mantle support??

After seeing the apu13 q&a i wonder why the devs puts up for dx so many years. Its crap and the debugging is happening in darkness. Hearing the devs i have to wonder what future dx have or if it does have a future in 5 years at all. It doesnt give basic functionality like eg how big a texture will be. All is going to Mantle and now i know its more from the quality of the tool than just pure speed.

As a consumer this is the worst time ever to buy an nv gfx as mantle will not only make your cgn faster but also give you more cpu power. Who can not use that in bf4? Nv will lower their prices as a response so things will be leveled. So better wait to buy nv if thats the preferred brand.

 

[OCGuy]

That doesn't refute what I said. Like I said. Go through the slides and you wouldn't need to have other people have to explain it to you.

Mantle is not proprietary. This guy is the 1st person to claim it is and I don't know how much he knows about anything tbh. After using Fudzilla as a source and dissing Ryan.

Mantle to power 15 Frostbite games; DICE calls for multi-vendor support

"Mantle yields considerable benefits in terms of development flexibility and performance, and it's worth implementing even in its current, vendor-locked state. Andersson wasn't the only developer to express a desire for multi-vendor support.

There's no telling yet whether Mantle will ever become a cross-vendor, cross-platform standard, or whether the future holds something different........"

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support

 

[Skurge]

Mantle to power 15 Frostbite games; DICE calls for multi-vendor support

"Mantle yields considerable benefits in terms of development flexibility and performance, and it's worth implementing even in its current, vendor-locked state. Andersson wasn't the only developer to express a desire for multi-vendor support.

There's no telling yet whether Mantle will ever become a cross-vendor, cross-platform standard, or whether the future holds something different........"

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support

The support is up to Nvidia and Intel when the SDK is released. That's what Johan means. Unless you think he doesn't believe anything in his slide. I think you didn't actually watch Johan during the keynote and instead just reading articles from other people who did, cause that is the only I see you making such a conclusion.

Even the 1st 3 points in the slide should tell you that. here.

Rebelion just announced Sniper Elite 3 and their future games will have support for Mantle. That confirms BF4, Thief, Starcitizen, Sniper Elite 3 and PvZ. That's 5 games already. Plus 15 more from EA through the next year. Sounds like it's gaining traction pretty quickly. Must be easy to implement. This should push Intel and Nvidia to support it, which would be better for ALL of PC gaming for everyone.

but whatever, I have no desire to go in circles. So make if it what you will, but don't claim in GCN only or it's proprietary when you clearly don't even understanding how it works.

 

[Gloomy]

Does Microsoft allow vendor-specific extensions in DirectX?

 

[krumme]

Mantle to power 15 Frostbite games; DICE calls for multi-vendor support

"Mantle yields considerable benefits in terms of development flexibility and performance, and it's worth implementing even in its current, vendor-locked state. Andersson wasn't the only developer to express a desire for multi-vendor support.

There's no telling yet whether Mantle will ever become a cross-vendor, cross-platform standard, or whether the future holds something different........"

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support

Andersson ofcource want multivendor support. It would make things easy for him.

But i bet he is not getting it.

The devs will then have to decide how many ressources they want to spend on new engines and aaa games on debugging old and slow dx crap to make it 20% faster for a minor minority of the discrete nv gfx market or they will go the cgn way for less ressources and better selling games.

It does look like they have decided what way to go.

Amd have no interest in letting the competitive performance edge slip they have with mantle. Nv will hate to bite half a mantle api without the tech to use it.

 

[3DVagabond]

So this is all about Mantle...which means you want to kill off all AMD competitors to realize this perfect world. I personally don't like monopolies.

Microsoft is more concerned with selling you their latest OS than they are in improving gaming. It's about time the developers and hardware manufacturers took control of their business rather than having M$ pull all of the strings.

 

[Meekers]

Even if Nvidia could get 95% of the benefits of Mantle without changing a thing I would be surprised if they went that route. I do not think they want to help quicken Mantle adoption. They would become reliant on AMD to play nice and keep Mantle working properly on their cards going forward.

Unless it becomes a truly open platform, out of AMD's control, I do not see Nvidia willingly getting involved.

 

[smackababy]

Strawmen and threadcrapping as usual from ocg and smbaby. As if Ryan didnt knew what he was talking about.

So, I will take this as a no then? There is absolutely no real demos or code examples showing the similarities of the APIs nor are there any demos or code examples showing Mantle actually performing better even in synthetic benchmarks?

Obviously, I'm the one threadcapping because I went on a broken English rant about Nvidia being bad and no real substance other than "Mantle is great because someone else said so."


I am done with this thread until the SDK comes out. The Mantle fanboys are rather silly.

 

[bystander36]

Microsoft is more concerned with selling you their latest OS than they are in improving gaming. It's about time the developers and hardware manufacturers took control of their business rather than having M$ pull all of the strings.

I don't have a problem with the dev's taking control, but don't act like MS hasn't tried to improve gaming. Even if they do it so they can sell their latest OS. If they weren't trying, OpenGL would rule the roost.

 

[3DVagabond]

There is still a abstraction layer. As long as GPU's have support for same features they should be able to run Mantle.

At least that's what the developers are saying.

This^^^ is what people who actually are working with Mantle say.

Providing the data helps the discussion. AMD is saying it "supports" other vendors, but functionality is limited to GCN. If it was just a driver, all AMD cards could use it. The slides say they "hope" vendors in the future will use it, meaning NV/Intel would have to invest money to get it to work, meaning it wont work.


http://vr-zone.com/articles/mantle-shows-promise-proprietary-nature-still-achilles-heel/64273.html

Mantle shows promise, but its proprietary nature is still its achilles heel





On a side note, the author takes a dig @ AT/Ryan Smith in the following paragraph due to the initial thoughts by basically everyone that Mantle would be in consoles as well as PC. I do not believe that is fair, as everyone was thinking the same thing at the time. Speculation always runs the risk of looking foolish later.....he is man enough to leave the article up. Did I just uncover this or am I late to the party?

I do hope that people click on the link and read the article. If they don't they might think the author has facts to back his position. Which he doesn't. He basically quotes what people at APU13, who actually understand Mantle, said and then states his skepticism with no facts to back it.

He proceeds to link to Fudzilla as the source of money paid by AMD to EA (Because Fudzilla are the shining example of journalistic integrity.) and then to Karmack (Who no longer develops games but is the expert on Mantle, I guess?) at an nVidia event as his reasons that Mantle won't catch on with other people. All pretty weak sauce.

 

[bystander36]

Let's just say a lot of different stuff has been said about Mantle. Everything from giving Dev's low level access to GCN, to being a general DirectX replacement that works for all brands.

The thing is, both can't be true. Some of these "facts" are not true.

If Mantle is a general API that could replace DirectX on all vendors, and AMD has done nothing to make it favor their architecture, then it would be pretty great for Dev's, and PC gaming in general, as it would allow for simpler development. They wouldn't have to rewrite code for all the different architectures.

If Mantle is to give AMD some huge advantage, then it obviously has architecture specific advantages, and you cannot expect other brands to jump on board, and all this talk is nonsense.

 

[smackababy]

and then to Karmack (Who no longer develops games but is the expert on Mantle, I guess?) at an nVidia event as his reasons that Mantle won't catch on with other people.

In Carmack's defense, he was developing games before Windows created the abstraction layer to the hardware and gave us DX. I think he is quite an expert when it comes to actually writing code.

 

[mariooi]

Recently, John Carmack spoke about AMD’s low level API, he believe that Sony and Microsoft will not be using the API for their respective next gen systems.

 

[frozentundra123456]

This^^^ is what people who actually are working with Mantle say.



I do hope that people click on the link and read the article. If they don't they might think the author has facts to back his position. Which he doesn't. He basically quotes what people at APU13, who actually understand Mantle, said and then states his skepticism with no facts to back it.

He proceeds to link to Fudzilla as the source of money paid by AMD to EA (Because Fudzilla are the shining example of journalistic integrity.) and then to Karmack (Who no longer develops games but is the expert on Mantle, I guess?) at an nVidia event as his reasons that Mantle won't catch on with other people. All pretty weak sauce.

I dont think anybody really knows yet how practical it is to use mantle on another platform.

There is really a lot of uncertainty over the entire process. Right now we have only the statements of developers (who are clearly not the most objective sources) with no hard data to back it up. There have also been a lot of contradictory statements. It was "Mantle is the console api" but well, no it isnt really, just similar. And for the entire time since it was introduced, the whole selling point was that it would work so efficiently on gcn cards because they were similar in architecture to the consoles. Now we get a statement, well, it will work with any gpu architecture that has the right features. They did not say what those features were though, so that is a pretty vague statement.

Intuitively, it does not seem logical to me that AMD would spend the time and resources to develop mantle if it could be easily adopted to intel and nVidia and give them the same advantages without their spending the time and money to develop it. But again, nobody really knows. In fact nobody even knows how much performance increase we will see on AMD hardware with mantle.

 

[3DVagabond]

Mantle to power 15 Frostbite games; DICE calls for multi-vendor support

"Mantle yields considerable benefits in terms of development flexibility and performance, and it's worth implementing even in its current, vendor-locked state. Andersson wasn't the only developer to express a desire for multi-vendor support.

There's no telling yet whether Mantle will ever become a cross-vendor, cross-platform standard, or whether the future holds something different........"

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support

This speaks to the possible multi vendor support. From the same article.

But the "pink elephant in the room," as he called it, is multi-vendor support. Andersson made it clear that, while it only supports GCN-based GPUs right now, Mantle provides enough abstraction to support other hardware—i.e. future AMD GPUs and competing offerings. In fact, Andersson said that most Mantle functionality can work on most modern GPUs out today. I presume he meant Nvidia ones, though Nvidia's name wasn't explicitly mentioned. In any event, he repeated multiple times that he'd like to see Mantle become a cross-vendor API supported on "all modern GPUs."

It's too early for any other vendors to get on board. Nothing has been defined. Would AMD pull the strings? What if nVidia or Intel wanted to add some functionality of their own? Would AMD allow it? Would thaat functionality then become open to all? Would their only be features that AMD supports allowed? Will it need to be licensed? Too many unanswered questions and we still don't have a working demonstration to go by.

 

[GaiaHunter]

I was one of the few that said Mantle wasn't being used in consoles around the time of the article and quite a few people were responding with "nut uh! Ryan says so".

The problem is that Ryan never said Mantle was being used in the consoles.
He said mantle is the same or very similar to the low level API in the Xbox.

Some people just aren't capable or don't want to understand the difference between saying Mantle is equal/similar to something in the consoles and saying that Mantle is being used in the consoles.

 

[0___________0]

Skurge is right about the differences in architectures being irrelevant. Mantle is supposed to be indifferent in that sense, like DX; and it's to AMD's benefit that it is. You have to remember that AMD won't be sitting on GCN 1.1 for the next 10 years like the consoles, that's why there is still an abstraction layer. The only thing stopping NV from creating the drivers and getting similar benefits out of Mantle is themselves.

I can't believe people are still convinced Ryan can do no wrong. He was completely wrong about his speculation about the XB1 being the source of Mantle, AMD has stated the PS4's API is closer to Mantle. We also know it was designed entirely from the ground up on and for PC, it is not a port of a console API. Ryan argued it was a direct port, maybe some people don't know what a port is; he thought 99% of your code on a console could be used on what he thought was the PC version of the API known as Mantle.

Anyways, I think the most important thing I have to say is this: I think AMD has already made it clear what they mean when they say "similar to consoles". Mantle exposes all functionality of the GPU and provides relatively low level access and incorporates HLSL for streamlined porting. Maybe some people expected you to re-use all the code you wrote for consoles, but that's not realistic. Mantle isn't going to stand or fall based on whether it's a copy from a console API, people shouldn't get so hung up over that.

 

[BoFox]

There is still a abstraction layer. As long as GPU's have support for same features they should be able to run Mantle.

At least that's what the developers are saying.

You have it wrong. The architecture does not need to be similar. It just needs to have the minimum feature set that Mantle requires. Like DX. Kepler may have the required features or it may not. We don't know. Or nVidia just writes and driver and the devs use which ever features work best on kepler. Theoretically making kepler run better with Mantle than DX. But at this point we aren't sure.

But devs and AMD have said it does not require GCN. That is why whatever the next architecture from AMD is, it will work with Mantle as it will have the same features as GCN plus more. Which Mantle will also make use of.

I say there is a probably a minimum set of features need for Mantle, that is probably why pre GCN cards do not support it. Obviously AMD can't say it will work with kepler cause they didn't build kepler so they wouldn't know if it would work or not. That's what I got from the slides and interviews anyway.

If you had paid read the slides or watched the videos you would know all of this.

No!!!! CUDA/PhysX is good, and Mantle is evil!!!

:biggrin: I joke!!