The AMD Mantle Thread

[krumme]

Yeah, its what im thinking, all layers moved to driver side, and the API is not low level at all and the API whiout the driver its useless, but i dont think there will be any diference on API(devs) side for different Mantle drivers, im sure the API will run on a compatible Mantle driver, making no diference for the devs on that part.
Now, diferent driver implementations to api funtions may yield diferent performance.

Excactly. Now how much different is xbox ps4 and mantle driver

At the same time we also go an explanation why its cutting cost to program for mantle.

Its what i have been saying all time. Mantle is not fragmentation. Its the oposite. Moving layers and complexity to the driver is only possible because the hardware pool is severely restricted. Of cource. There is no free meal here.

That said i have no doubt the mantle solition is the right cut between driver and api. We know why the developers love it - less work lol.

 

[VulgarDisplay]

The amount of hype coming from these developers leads me to believe that Mantle is going to be awesome...

Really, since there are only 3 GPU vendors these days, it wouldn't be impossible to just have 3 low level api's. Not like back when Glide was around and there were twice as many GPU companies.

 

[SiliconWars]

I can't believe that every single major developer isn't coding Mantle support into their engines, right now. 2 "man months" is nothing. Anyone not doing it is endangering their engine.

 

[Abwx]

So much for the doom and gloom speculations built
on alleged added costs for Mantle..

 

[krumme]

I can't believe that every single major developer isn't coding Mantle support into their engines, right now. 2 "man months" is nothing. Anyone not doing it is endangering their engine.

If oxide 2 man monts is just remotely true or close to realitety mantle is not only comming everywhere. But the adoption is going to be within 2014!

2 man monts is an absurdly low cost. Its hardly measurable.

How i read apu13 mantle is there and amd is starting to gain support for hsa. The leverage is mobiles and their java work. Its going to be far more tough to gain support and they need fully hsa enabled consoles within 2-3 years to get some traction.

But mantle is there to stay and very much present.

 

[krumme]

Seeing the oxide presentation especially the "you know its bad when..." is an eyeopener for me.

I wouldnt have guessed call problems hit the artistic side so easily with the prior dx solution. "More elements in the same texture" damn. No wonder they are adopting mantle necause dx is just plain bad.

But what is also aparent is dx specific optimizations will get far lower in the future. Fine for me, what a waste of ressources and talent on next to nothing.
Let that old dog die fast.

 

[blastingcap]

If oxide 2 man monts is just remotely true or close to realitety mantle is not only comming everywhere. But the adoption is going to be within 2014!

2 man monts is an absurdly low cost. Its hardly measurable.

How i read apu13 mantle is there and amd is starting to gain support for hsa. The leverage is mobiles and their java work. Its going to be far more tough to gain support and they need fully hsa enabled consoles within 2-3 years to get some traction.

But mantle is there to stay and very much present.

Just because something is technically superior does not mean it will prevail in the marketplace. Sweeney is apparently deep in bed with NVidia and he also controls Unreal, one of the, if not THE most popular game engine out there. It was unreal how much he kept talking about Nvidia, Nvidia, Nvidia throughout the entire talk. You've got to get people like him on board. I'm not sure AMD has enough clout to do that.

 

[3DVagabond]

Seeing the oxide presentation especially the "you know its bad when..." is an eyeopener for me.

I wouldnt have guessed call problems hit the artistic side so easily with the prior dx solution. "More elements in the same texture" damn. No wonder they are adopting mantle necause dx is just plain bad.

But what is also aparent is dx specific optimizations will get far lower in the future. Fine for me, what a waste of ressources and talent on next to nothing.
Let that old dog die fast.

I could see optimizations that are designed to reduce calls going away. From what they say, they only reduce quality.

 

[Carfax83]

Ryan Shrout - "Oxide claims that it only took 2 "man months" to support Mantle. "Not much work or expensive" to integrate it."

"API overhead reduced by at least 10x. In some cases StarSwarm is 3x faster with Mantle."

"Completely changes the landscape for multi-core. Claims an FX-8350 is now just as fast as a Core i7-4770K"

James Prior - "With Mantle, an @AMDFX 8350 underclocked to 2GHZ, the game is still GPU bound on a R9 290X - no loss in frame rate"

I would call that some staggering info.

If that's true, NVidia really needs to jump on this as well. Being CPU bound in games (looks at BL2) is really a pain in the ass...

 

[Abwx]

Just because something is technically superior does not mean it will prevail in the marketplace. Sweeney is apparently deep in bed with NVidia and he also controls Unreal, one of the, if not THE most popular game engine out there. It was unreal how much he kept talking about Nvidia, Nvidia, Nvidia throughout the entire talk. You've got to get people like him on board. I'm not sure AMD has enough clout to do that.

Once the competition get flashy games we ll see how the rest
react when witnessing their games as being grossly shaped
items in comparison of micrometrics polished master pieces....

 

[3DVagabond]

Just because something is technically superior does not mean it will prevail in the marketplace. Sweeney is apparently deep in bed with NVidia and he also controls Unreal, one of the, if not THE most popular game engine out there. It was unreal how much he kept talking about Nvidia, Nvidia, Nvidia throughout the entire talk. You've got to get people like him on board. I'm not sure AMD has enough clout to do that.

There's one in every crowd

He's not more influential than the entire collective.

 

[Carfax83]

Just because something is technically superior does not mean it will prevail in the marketplace. Sweeney is apparently deep in bed with NVidia and he also controls Unreal, one of the, if not THE most popular game engine out there. It was unreal how much he kept talking about Nvidia, Nvidia, Nvidia throughout the entire talk. You've got to get people like him on board. I'm not sure AMD has enough clout to do that.

If the advantages of implementing Mantle are as great as being suggested, they (EPIC and NVidia) won't really have a choice because consumers will demand it, and it will make NVidia look bad in benchmarks.

It could be a huge game changer for PC gaming. But I would not underestimate Microsoft. DX12 could be something like Mantle, as I know they've expressed great interest in lowering CPU overhead due to the API.

 

[blastingcap]

There's one in every crowd

He's not more influential than the entire collective.

He might be pressured into it if everyone else piles on, but it's clear that he will not lead the charge. Unreal is in a LOT of games. http://en.wikipedia.org/wiki/List_of_Unreal_Engine_games

Don't get me wrong, I want Mantle to succeed, but I've seen plenty of technically superior things lose in the marketplace anyway.

If the advantages of implementing Mantle are as great as being suggested, they (EPIC and NVidia) won't really have a choice because consumers will demand it, and it will make NVidia look bad in benchmarks.

It could be a huge game changer for PC gaming. But I would not underestimate Microsoft. DX12 could be something like Mantle, as I know they've expressed great interest in lowering CPU overhead due to the API.

You have a very flattering view of gamers. I tend to think of gamers as average people who buy whatever is shoveled onto them via massive marketing dollars. See, e.g., the COD franchise.

Mantle is likely not SO much of a leap that it allows companies supporting it to make things SO much better looking or smoother that it becomes an irresistible force. There are exceptions, though, such as massive RTS genres which are in dire need of a fix, as Oxide said. I have already vowed to avoid all Creative Assembly titles until they better multithread their games, for instance, and that's something Mantle could help with. But other genres I don't think would benefit as much. As Johan from EA said, even if you alleviate some bottleneck with Mantle so the game would run 4x faster, there are a tons of other bottlenecks, some of which are not improved by Mantle, so your actual gain isn't 4x, it might just be 10% or something.

I'm hoping Valve will work closely with AMD on bringing Mantle to Steambox titles (utilizing Valve's clout with developers, even small time ones count if you add them all up) and thus make PC hardware more cost-competitive with heavily-subsidized consoles (XBO, PS4, Wii U).

 

[Barbatruuk]

What does one have to do with the other?

Well, I was just thinking. A lot of people are waiting out for the aftermarket 290's and I guess there'd be quite a few people undecided whether to go for aftermarket AMD or Nvidia cards.

What if AMD gets the 'doubters' excited enough along the road to BF4 Mantle patch and releases the aftermarket 290's to coincide with the BF4 Mantle patch (or perhaps a tad sooner), the doubters could be going AMD.

It was just a thought, as we STILL don't know the release date for aftermarket cards :\

I, for sure, am quite excited about the released info on Mantle (would've loved some actual BF4 performance numbers though).
Can't wait to get my hands on an aftermarket 290 and play Mantle BF4 on my invigorated fx-8350

 

[Atreidin]

Ryan Shrout - "Oxide claims that it only took 2 "man months" to support Mantle. "Not much work or expensive" to integrate it."

"API overhead reduced by at least 10x. In some cases StarSwarm is 3x faster with Mantle."

"Completely changes the landscape for multi-core. Claims an FX-8350 is now just as fast as a Core i7-4770K"

James Prior - "With Mantle, an @AMDFX 8350 underclocked to 2GHZ, the game is still GPU bound on a R9 290X - no loss in frame rate"

I would call that some staggering info.

With information like that, it would be stupid to leave Mantle out of a game engine. Low cost, much better performance, and opening up your market to more low-powered PCs sounds like a big win. You wouldn't need to sell many extra copies to make up for the development cost. (I'm waiting for the "But AMD is secretly paying them all to do it" mantra being repeated by people who like saying wild speculation as authoritative statements.)

 

[Barbatruuk]

With information like that, it would be stupid to leave Mantle out of a game engine. Low cost, much better performance, and opening up your market to more low-powered PCs sounds like a big win. You wouldn't need to sell many extra copies to make up for the development cost. (I'm waiting for the "But AMD is secretly paying them all to do it" mantra being repeated by people who like saying wild speculation as authoritative statements.)

:biggrin:

 

[moonbogg]

Well, I was just thinking. A lot of people are waiting out for the aftermarket 290's and I guess there'd be quite a few people undecided wether to go for aftermarket AMD or Nvidia cards.

What if AMD gets the 'doubters' excited enough along the road to BF4 Mantle patch and releases the aftermarket 290's to coincide with the BF4 Mantle patch (or perhaps a tad sooner), the doubters could be going AMD.

It was just a thought, as we STILL don't know the release date for aftermarket cards :\

I, for sure, am quite excited about the released info on Mantle (would've loved some actual BF4 performance numbers though).
Can't wait to get my hands on an aftermarket 290 and play Mantle BF4 on my invigorated fx-8350

You just might not be kidding. I can easily see that if Mantle actually takes off and gives gamers a lot, the price of FX chips could go up IMO, and they would be worth a little extra if Mantle rocks. Just the name alone and the fact that AMD owns Mantle and the chips, plus the consoles etc, I can see a surprisingly bright future for AMD hardware (long shot but I can see it).
You see, I am of the opinion that gamers made all of these tech companies what they are today. Without the passion and excitement of gamers driving hardware advancements, there wouldn't have ever been the kind of advancement that we have seen over the last few decades.
I think that if AMD gets gamers excited again that it will be gamers that save AMD and return it to greatness. Nothing else can do it.

 

[VulgarDisplay]

You just might not be kidding. I can easily see that if Mantle actually takes off and gives gamers a lot, the price of FX chips could go up IMO, and they would be worth a little extra if Mantle rocks. Just the name alone and the fact that AMD owns Mantle and the chips, plus the consoles etc, I can see a surprisingly bright future for AMD hardware (long shot but I can see it).
You see, I am of the opinion that gamers made all of these tech companies what they are today. Without the passion and excitement of gamers driving hardware advancements, there wouldn't have ever been the kind of advancement that we have seen over the last few decades.
I think that if AMD gets gamers excited again that it will be gamers that save AMD and return it to greatness. Nothing else can do it.

Gamers are to computers what porn is to video standards.

 

[Barbatruuk]

You just might not be kidding. I can easily see that if Mantle actually takes off and gives gamers a lot, the price of FX chips could go up IMO, and they would be worth a little extra if Mantle rocks. Just the name alone and the fact that AMD owns Mantle and the chips, plus the consoles etc, I can see a surprisingly bright future for AMD hardware (long shot but I can see it).
You see, I am of the opinion that gamers made all of these tech companies what they are today. Without the passion and excitement of gamers driving hardware advancements, there wouldn't have ever been the kind of advancement that we have seen over the last few decades.
I think that if AMD gets gamers excited again that it will be gamers that save AMD and return it to greatness. Nothing else can do it.

Yeah, I really hope AMD can pull this all off. For the sake of better and faster graphics, better console ports, older hardware still usable for current games while high end hardware gets some extra grunt.

 

[selni]

Is this true that John wants Nvidia to adopt mantle.

If it's even possible for nvidia to run mantle code on kepler efficiently then this got a whole lot more interesting.

I don't think a single vendor proprietary platform is a great direction for the industry to go in, but if this is a general directx replacement that's *much* bigger news.

 

[moonbogg]

If it's even possible for nvidia to run mantle code on kepler efficiently then this got a whole lot more interesting.

I don't think a single vendor proprietary platform is a great direction for the industry to go in, but if this is a general directx replacement that's *much* bigger news.

Nvidia would never be able to see past their pride in such a way. They'd never use AMD's Mantle, no matter how awesome it would make Nvidia GPU's perform. Every single Nvidia fan could scream for it, and they would never do it. They would let their company die before doing that.

 

[3DVagabond]

Gamers are to computers what porn is to video standards.

Are you saying that you think Mantle will help my pr0n?

 

[beginner99]

They also just said that Mantle is not tied to AMD and was designed not only for GCN. He just said Mantle would be more efficient for "the other vendor"

Mantle is an API and like any API it can have different implementations. In AMDs image this is clearly named as "Mantle Driver", a layer below the Mantle API.

It is just like in Java where Java = Mantle API and the JVM = Mantle Driver.

And I was correct even so AMD haters denied it.

 

[3DVagabond]

If it's even possible for nvidia to run mantle code on kepler efficiently then this got a whole lot more interesting.

I don't think a single vendor proprietary platform is a great direction for the industry to go in, but if this is a general directx replacement that's *much* bigger news.

Just because it's not specifically GCN only, doesn't mean that Kepler can run it. It just might simply mean that AMD can continue to advance their architecture beyond GCN and maintain Mantle compatibility, and that other companies could design Mantle compatible hardware.

 

[Skurge]

If that's true, NVidia really needs to jump on this as well. Being CPU bound in games (looks at BL2) is really a pain in the ass...

Damn, and I just ordered an i5. I knew I should have gotten an i7. oh well. We will see what actual numbers are.