The AMD Mantle Thread

[desprado]

From your link:

go search bro
Respawn is Nvidia Partners

 

[desprado]

It is sponsor heavily by Nvidia
https://www.facebook.com/NVIDIAGeForce/posts/672908259395301

 

[VulgarDisplay]

It is more likely to be 20mn and check Google and all most all rumors are saying Nvidia is developing 20mn cards.

Yes they are, but they aren't coming out in Q1. More like Q3 at the earliest.

 

[krumme]

Listen, lets not make this like a cheerleading football game. That is not what i was doing. I'm simply stating a fact. To me, as of this moment, Mantle isn't any better than the deceased "Stream" API AMD announced with bells on way back when. AMD tends to falter at the finish line or not follow through.
Stream
Bullet
Coolers
Mantle?

not saying Nvidia doesn't make it's own mistakes. Not at all.
It is what it is.

I tend to agree here. Add amd is financially very constrained. I dont think most people here know how that drags down developing a company compared to beeing the big guy. Its a staggerimg difference. Otellini knew and he used that. Using your cash the outmaneuver compettion is very effective i tell you as you can use the cash everywhere and get effect.

What differs with mantle to the above is it builds on the technical competences in amd and not the more marketing like decisions and competences. Secondly from a broader perspective the investment is more or less paid - all the hardware ip and consoles is there so to speak.

I am sure nv will come up with something. It doesnt have to be like mantle. Who gives. They have loads of failires behing them but also hogly successfull solutions. They are a very innovative company on the business side. Very underrated in these forums btw. But its where the big money is.

 

[krumme]

AMD would have debuted the cards at higher prices if that were the case, they did know what Mantle is bring to the table.

Rory Read is after a much bigger fish - breaking the back of Nvidia's consumer GPU business on a cost performance rack.

AMD is currently a cost/performance leader across the AIB market even after Nvidia's initial price reductions, what margins will be left if Nvidia has to contend with a further 25-35% Mantle induced cost/performance deficit?

What happens when Kaveri is released able to take advantage of the full HSA/hUMA APU console code? Kaveri + Mantle will functionally duplicate the console hardware.

What happens to prices after Nvidia is broken on the price/performance rack is a different matter.

Nv brand is worth 25%. Look at the prices people gladly pay for 25% worse perf/$. Nv is clearly competitive even if mantle adds 20% to the speed. Besides they have their cash cow on the pro market. Amd got none. Only the consoles looks to generate some solid profit.

 

[ShintaiDK]

Its amazing the confidence and such in a product that havent been demoed, no performance estimates. And only coming to a handfuld of games the next couple of years.

Sofar the only companies that likes Mantle is the ones AMD pays. And thats not a surprise due to the minority segment.

Maybe people should wait to we actually see performance. Because the hype about it is at a terrible level that can only go wrong. Consoles didnt have anything to do with Mantle as first predicted for example.

 

[MeldarthX]

Its amazing the confidence and such in a product that havent been demoed, no performance estimates. And only coming to a handfuld of games the next couple of years.

Sofar the only companies that likes Mantle is the ones AMD pays. And thats not a surprise due to the minority segment.

Maybe people should wait to we actually see performance. Because the hype about it is at a terrible level that can only go wrong. Consoles didnt have anything to do with Mantle as first predicted for example.

We'll see; but you so sure its only companies AMD's pays? I've love to see the proof on this; we've gotten leaks of some of the big boys that have signed on....better yet; we'll know this week....

We are going to have some demos of it; handful of games? well 15 games with FB3 - one of the biggest games next year; Thief 4; along with true audio in that game.

Full RTS game engine for stardock... that's two engines; Star Cit will use it; which means either Chris is baking it into the game; or its getting baked in by Crytek.....*most likely both* Activition is using it; I can say other big companies will be there also. We'll see just how much of a game changer this is....specially with BF4 mantle patch supposed to be coming sooner than thought also....

 

[ShintaiDK]

We'll see; but you so sure its only companies AMD's pays? I've love to see the proof on this; we've gotten leaks of some of the big boys that have signed on....better yet; we'll know this week....

We are going to have some demos of it; handful of games? well 15 games with FB3 - one of the biggest games next year; Thief 4; along with true audio in that game.

Full RTS game engine for stardock... that's two engines; Star Cit will use it; which means either Chris is baking it into the game; or its getting baked in by Crytek.....*most likely both* Activition is using it; I can say other big companies will be there also. We'll see just how much of a game changer this is....specially with BF4 mantle patch supposed to be coming sooner than thought also....

Thief is another example of a gaming evolved, aka AMD payed. Star Citizen supports everything that pays it money, including PhysX. But I still consider that title vapourware until it gets closer. Right now it simply slides on its timeline. Is Activision using it? The only comment they had sofar was about a wrapper on twitter.

So lets wait and see honestly.

And because an engine supports it, doesnt mean it will be ingame. Just like PhysX.

The biggest problem for developers, is to justify to code a second renderpath for a segment that is below 10% marketshare, when AMD doesnt cover the cost. Sofar AMD payed all the titles we see. No developer did it for their own money.

 

[3DVagabond]

We'll see; but you so sure its only companies AMD's pays? I've love to see the proof on this; we've gotten leaks of some of the big boys that have signed on....better yet; we'll know this week....

We are going to have some demos of it; handful of games? well 15 games with FB3 - one of the biggest games next year; Thief 4; along with true audio in that game.

Full RTS game engine for stardock... that's two engines; Star Cit will use it; which means either Chris is baking it into the game; or its getting baked in by Crytek.....*most likely both* Activition is using it; I can say other big companies will be there also. We'll see just how much of a game changer this is....specially with BF4 mantle patch supposed to be coming sooner than thought also....

They're supposed to demo BF4 using Mantle. It's only a few more days. Judging by what the devs and AMD are saying, it should be a win for AMD, a win for the developers, and a win for gamers. What's the reason not to be hopeful?

 

[3DVagabond]

Sofar AMD payed all the titles we see. No developer did it for their own money.

i haven't seen a source for that. Do you have one?

 

[desprado]

Its amazing the confidence and such in a product that havent been demoed, no performance estimates. And only coming to a handfuld of games the next couple of years.

Sofar the only companies that likes Mantle is the ones AMD pays. And thats not a surprise due to the minority segment.

Maybe people should wait to we actually see performance. Because the hype about it is at a terrible level that can only go wrong. Consoles didnt have anything to do with Mantle as first predicted for example.

of course no dev are stupid investing on a company which dont even have GPU or CPU market majority or share unless the company it self pay them to use.

 

[Minkoff]

of course no dev are stupid investing on a company which dont even have GPU or CPU market majority or share unless the company it self pay them to use.

Well, let's agree to disagree.
1. When you make a strategy you put goals ahead --> trying to gain marketshare.
2. The investment (as you point) is not for Devs to make but rather AMD. Mantle is a long term investment and aims at providing additional free perf over the refresh/new arch gpus and, ofcourse it is different from Nvidia twimtbp since it focuses on engines not per game development.

To constantly point the AMD paid for the use of Mantle in Battlefileld 4 is also wrong, since we do not know what that payment is for. If Mantle is a brainchild of both DICE and AMD it does not make much sence for AMD to pay for the use of it now, does it?

Edit:

i haven't seen a source for that. Do you have one?

I agree with you. And there is always more than one logic to things.
If low-level-console type development for PC has been requested from developers...I don't see the payment theory beeing true. Not saying that there isn't any...

 

[Keysplayr]

They're supposed to demo BF4 using Mantle. It's only a few more days. Judging by what the devs and AMD are saying, it should be a win for AMD, a win for the developers, and a win for gamers. What's the reason not to be hopeful?

And also, when that happens and if and when Mantle provides a substantial performance boost in BF4, that is great, but it is also JUST BF4.
I don't want to see sweeping assumptions that EVERY game will be exactly the same.
Conversely, if the performance improvement is minimal or non-existent, I don't want to start hearing, "Wait til the next game comes out, THAT one will be the one!! "
Nuh uh.
Game by game basis is the most intelligent approach in discussing anything to do with Mantle. IMHO.

 

[redzo]

Still you have to face the "green" problem, since mantle runs on AMD only, you still have to fully support directx with good quality code. And then you've got Microsoft who is not going to allow mantle on xbox because it will challenge directx's dominance. So the entire cost reduction argument might go to waste: supporting two different api's(dx + mantle) is not going to be cheaper than supporting one api(dx only).

Even if mantle will bring a miraculous 15-20% performance gain across the board, no dev is going to adopt it(unless you pay them) because it increases development costs/hell. I think at the end of the day, devs will just say: "Not another 3d api. You've already got dx and opengl. The fact that it's vendor specific(amd only), makes it pointless".

Just my humble opinion. Time will tell.

 

[Keysplayr]

i haven't seen a source for that. Do you have one?

I haven't either. Would appreciate a source and be assured I'm not just listening to someone's "feeling" on this.
Thanks.

 

[desprado]

Still you have to face the "green" problem, since mantle runs on AMD only, you still have to fully support directx with good quality code. And then you've got Microsoft who is not going to allow mantle on xbox because it will challenge directx's dominance. So the entire cost reduction argument might go to waste: supporting two different api's(dx + mantle) is not going to be cheaper than supporting one api(dx only).

Even if mantle will bring a miraculous 15-20% performance gain across the board, no dev is going to adopt it(unless you pay them) because it increases development costs/hell. I think at the end of the day, devs will just say: "Not another 3d api. You've already got dx and opengl. The fact that it's vendor specific(amd only), makes it pointless".

Just my humble opinion. Time will tell.

Agree but one point more is that IMO that Mantle will give full performance over there APU and of course they try gain there dead shares in CPU market and most of here anandtech member have intel processor.

 

[GaiaHunter]

Still you have to face the "green" problem, since mantle runs on AMD only, you still have to fully support directx with good quality code. And then you've got Microsoft who is not going to allow mantle on xbox because it will challenge directx's dominance. So the entire cost reduction argument might go to waste: supporting two different api's(dx + mantle) is not going to be cheaper than supporting one api(dx only).

Even if mantle will bring a miraculous 15-20% performance gain across the board, no dev is going to adopt it(unless you pay them) because it increases development costs/hell. I think at the end of the day, devs will just say: "Not another 3d api. You've already got dx and opengl. The fact that it's vendor specific(amd only), makes it pointless".

Just my humble opinion. Time will tell.

Microsoft doesn't have to allow or not mantle in the xbone.
Mantle can be included in the game engine - not that it will have to be.

Mantle just has to be similar to the xbone low level constructs and since they both target the same hardware architecture it shouldn't be to hard.

 

[Yarn]

Thief is another example of a gaming evolved, aka AMD payed. Star Citizen supports everything that pays it money, including PhysX. But I still consider that title vapourware until it gets closer. Right now it simply slides on its timeline. Is Activision using it? The only comment they had sofar was about a wrapper on twitter.

So lets wait and see honestly.

And because an engine supports it, doesnt mean it will be ingame. Just like PhysX.

The biggest problem for developers, is to justify to code a second renderpath for a segment that is below 10% marketshare, when AMD doesnt cover the cost. Sofar AMD payed all the titles we see. No developer did it for their own money.

I'm going to ask you one more time where are you getting this <10% from?

 

[Keysplayr]

ShintaiDK, you really need to start posting sources for your comments. Even I am thinking you are just making this stuff up and most people find it easier to call BS than to look it up for themselves.

 

[ShintaiDK]

I'm going to ask you one more time where are you getting this <10% from?

Steam hardware survey.

 

[Noctifer616]

https://twitter.com/FrostbiteEngine/status/399882045020987392

There is a Mantle presentation at the Frostbite Dev days 13 in Stockhol.

There is no official stream or VOD for this event. Does anyone have any info on it or do we have to wait a few more days (should be the same presentation as the one on the AMD summit)?

 

[ShintaiDK]

ShintaiDK, you really need to start posting sources for your comments. Even I am thinking you are just making this stuff up and most people find it easier to call BS than to look it up for themselves.

You mean the source that shows that Thief and BF4 for example is an AMD sponsored game via the Gaming Envolved sponsorship? Its no different than the crap nVidia does to pay companies to use PhysX.

http://sites.amd.com/us/game/games/Pages/games.aspx

 

[ShintaiDK]

Microsoft doesn't have to allow or not mantle in the xbone.
Mantle can be included in the game engine - not that it will have to be.

Mantle just has to be similar to the xbone low level constructs and since they both target the same hardware architecture it shouldn't be to hard.

You cant run a driver without MS permission. So you cant run Mantle in any way on Xbox One without MS giving the ok. And it would then have to be a system update to enable it.

 

[Silverforce11]

It's interesting to see the polarity amongst the camps here..

How can you be so sure Mantle will be an awesome success as claimed? You cannot. We don't even have BF4 performance results yet. It's all hype.

Likewise, how can you be so sure Mantle will fail hard, to justify the endless trolling going on here? You cannot, you have NOTHING to substantiate that it wont perform well.

So really, take a chill pill and wait and see. It's interesting times ahead.

 

[GaiaHunter]

You cant run a driver without MS permission. So you cant run Mantle in any way on Xbox One without MS giving the ok. And it would then have to be a system update to enable it.

Is physX supported by DX and MS?