The AMD Mantle Thread

[Keysplayr]

Slide 15 - http://www.frostbite.com/connect/#battlefield-4-frostbite-mantle



Says who?

It's just what any dev would ask for. Yet another API to worry over than only benefits the minority of graphics card owners. Yeah, I bet that happens on a daily basis.

 

[SiliconWars]

It's just what any dev would ask for. Yet another API to worry over than only benefits the minority of graphics card owners. Yeah, I bet that happens on a daily basis.

Nvidia were given every chance. They didn't take it.

AMD is the majority btw, but I'm sure we've been over that one before.

I noticed you deflected instead of answering the question however. Something you do a lot.

 

[ShintaiDK]

Its amazing AMD had to pay DICE 8mio$ for something DICE asked for. That really shows confidence there.

 

[krumme]

Perhaps nv will take take the chance and support mantle.

 

[SiliconWars]

Its amazing AMD had to pay DICE 8mio$ for something DICE asked for. That really shows confidence there.

Almost as amazing as Nvidia paying $5 million for a game that's 7-8 months out.

 

[BallaTheFeared]

Perhaps nv will take take the chance and support mantle.

Yes all they'd have to give up is their superior HPC uarch.

 

[Erenhardt]

Its amazing AMD had to pay DICE 8mio$ for something DICE asked for. That really shows confidence there.

Did they pay DICE, or will it cost them that much? There is a difference.

 

[krumme]

Yes all they'd have to give up is their superior HPC uarch.

Well i dont beliewe it either. Right now it seems for mantle the driver is very thin and thats the purpose of it. Then ms and sony adds their own version of mantle api on top of it. Probably sharing a lot of building blocks with ms probably taking the more dx route. They can label it whatever.
If amd wants quick success for mantle it needs to open it. I dont know if any bytes and bits could be used by nv in some form?. Or intel for that matter. Just a thought.

 

[krumme]

Did they pay DICE, or will it cost them that much? There is a difference.

We will see if this difference matters.
Imho it seems like a very low amount but thats probably because amd dont have more cash at hand even with an opportunity like this? Lol.

 

[Erenhardt]

We will see if this difference matters.
Imho it seems like a very low amount but thats probably because amd dont have more cash at hand even with an opportunity like this? Lol.

I mean, AMD could pay DICE to do mantle for them, or AMD could have their own engineers make mantle, which would tie to staff cost. Or a combination of both

 

[OCGuy]

Perhaps nv will take take the chance and support mantle.

huh....

 

[OCGuy]

Nvidia were given every chance. They didn't take it.

AMD is the majority btw, but I'm sure we've been over that one before.

Have you read anything that is going on with mantle? It is not in consoles, and Intel by far leads PC graphics, with NV leading discreet.

 

[VulgarDisplay]

I don't think anyone has thought of this yet, but here goes...

What if we all have our mantle opinions backwards. What if a game coded for mantle is easily portable to low level API's on PS4 and Xbone. So dev codes for PC first using Mantle, and it is an easy port to the PS4 and Xbone low level API's from there?

AMD has full access to the hardware on both consoles. Is there a reason why we couldn't all have it backwards in our speculation so far?

That would explain why they are talking about how it will decrease development time, and cost.

 

[bystander36]

I don't think anyone has thought of this yet, but here goes...

What if we all have our mantle opinions backwards. What if a game coded for mantle is easily portable to low level API's on PS4 and Xbone. So dev codes for PC first using Mantle, and it is an easy port to the PS4 and Xbone low level API's from there?

AMD has full access to the hardware on both consoles. Is there a reason why we couldn't all have it backwards in our speculation so far?

That would explain why they are talking about how it will decrease development time, and cost.

Most the stuff I keep reading, talk about being able to use parts of the console code on PC Mantle, so i don't think this is the case.

And the idea of Mantle decreasing development time is ridiculous. The only way it could reduce dev time is if they ignored non GCN cards (the vast majority of PC users).

 

[SiliconWars]

I don't think anyone has thought of this yet, but here goes...

What if we all have our mantle opinions backwards. What if a game coded for mantle is easily portable to low level API's on PS4 and Xbone. So dev codes for PC first using Mantle, and it is an easy port to the PS4 and Xbone low level API's from there?

AMD has full access to the hardware on both consoles. Is there a reason why we couldn't all have it backwards in our speculation so far?

That would explain why they are talking about how it will decrease development time, and cost.

Well you're not the first, but you're correct.

http://forum.beyond3d.com/showpost.php?p=1794610&postcount=217

I'm not 100% convinced about this part of it yet, but I can see how it would be a benefit.

 

[SiliconWars]

And the idea of Mantle decreasing development time is ridiculous. The only way it could reduce dev time is if they ignored non GCN cards (the vast majority of PC users).

Think of it this way.

AMD has a large GCN graphics base of millions. Let's pretend that instead of being a PC, it's a console called "Mental".

While developing a game for one of the consoles, the devs realise that porting it to "Mental" would be a fairly simple job that brought them extra revenue from the millions of "Mental" owners.

Soooo...why wouldn't they do that? They already did it for the PS4, the XBOX and the Wii right?

 

[Teizo]

Think of it this way.

AMD has a large GCN graphics base of millions. Let's pretend that instead of being a PC, it's a console called "Mental".

While developing a game for one of the consoles, the devs realise that porting it to "Mental" would be a fairly simple job that brought them extra revenue from the millions of "Mental" owners.

Soooo...why wouldn't they do that? They already did it for the PS4, the XBOX and the Wii right?

Did you purposely write "Mental" rather than "Mantle"?

 

[bystander36]

Think of it this way.

AMD has a large GCN graphics base of millions. Let's pretend that instead of being a PC, it's a console called "Mental".

While developing a game for one of the consoles, the devs realise that porting it to "Mental" would be a fairly simple job that brought them extra revenue from the millions of "Mental" owners.

Soooo...why wouldn't they do that? They already did it for the PS4, the XBOX and the Wii right?

Even if Mantle could port 100% to the consoles, it still would not help development time one bit, as it would have been the same amount of work if they just wrote it for the console. However, you even recognize that it would require some extra work to port it. If that was all that was required, then it would speed up development, as it would eliminate the need to write a DX code path, but because most PC users do not have a Mantle capable card, they still have to write for DX.

That means there is extra work being done to support Mantle. Because they cannot ignore DX users, Mantle just becomes an additional code path to support. It is not a code path in place of anything, as such, there is absolutely no saving in time.

 

[desprado]

Nvidia were given every chance. They didn't take it.

AMD is the majority btw, but I'm sure we've been over that one before.

I noticed you deflected instead of answering the question however. Something you do a lot.

can u post ur source say that AMD has the majority.If u say i can show 4 or 6 site which say Nvidia has 65% of the Pc user.

 

[Keysplayr]

can u post ur source say that AMD has the majority.If u say i can show 4 or 6 site which say Nvidia has 65% of the Pc user.

This guy is including APU's.

 

[VulgarDisplay]

Even if Mantle could port 100% to the consoles, it still would not help development time one bit, as it would have been the same amount of work if they just wrote it for the console. However, you even recognize that it would require some extra work to port it. If that was all that was required, then it would speed up development, as it would eliminate the need to write a DX code path, but because most PC users do not have a Mantle capable card, they still have to write for DX.

That means there is extra work being done to support Mantle. Because they cannot ignore DX users, Mantle just becomes an additional code path to support. It is not a code path in place of anything, as such, there is absolutely no saving in time.

What we are trying to say is that AMD will likely have a way for dev's to code for mantle on the PC and then easily port to the PS4 and the Xbone. So, code once for PC, get your highly optimized xbone and PS4 at one time. Then code for DX. So instead of having to code for both consoles you code for Mantle and PC and get both your console ports at one time.

That's really the only way that this could work where I see it actually saving dev's money.

 

[frozentundra123456]

What we are trying to say is that AMD will likely have a way for dev's to code for mantle on the PC and then easily port to the PS4 and the Xbone. So, code once for PC, get your highly optimized xbone and PS4 at one time. Then code for DX. So instead of having to code for both consoles you code for Mantle and PC and get both your console ports at one time.

That's really the only way that this could work where I see it actually saving dev's money.

It is all speculation that they will code for PC first. No matter how great mantle is on PC, the vast majority of money will be made in consoles, so my best guess would be that they would still code for consoles first. At least that would be my expectation until developers clearly state otherwise.

The problem for developers is that they still have to code for 2 separate codepaths, and probably will sell very few more games for PC than if they only had to code for DX. What I mean is if mantle shows a good performance improvement, I doubt that will result in more PC versions being sold, just a shift of the market share of graphics cards, which developers dont care about, unless one of them is pumping them loads of cash.

 

[krumme]

It is all speculation that they will code for PC first. No matter how great mantle is on PC, the vast majority of money will be made in consoles, so my best guess would be that they would still code for consoles first. At least that would be my expectation until developers clearly state otherwise.

The problem for developers is that they still have to code for 2 separate codepaths, and probably will sell very few more games for PC than if they only had to code for DX. What I mean is if mantle shows a good performance improvement, I doubt that will result in more PC versions being sold, just a shift of the market share of graphics cards, which developers dont care about, unless one of them is pumping them loads of cash.

Good argument and surely there will be change of marketshare. But i think ea look at it differently - and ofcource they dont care about nv nor amd.
For them its about taking the leasure time and money. Games is competing with other games but also eg. Movies, other hobbies, tv, surfing time (money).
Notice it. Mantle is a way to enlargen their current marketshare by providing a tool for the future that enables next level og gaming experience. Thats why dice on a lower strstegic level talk eg about multi gpu vr.
The 8m for dice next gen engine launched with bf4 is the same as zero imho. Look at it this way; what about all 3 engines for 18m? Who wouldnt pay that?

 

[krumme]

This guy is including APU's.

And in this situation its the right thing to do. Nobody who sell games cares if its discrete or integrated. All the consoles is apus. And probably the apus with weaker cpu gets the highest benefit because of reduced drawcalls. That means means you have a larger market for your games. All those craptops with 4 core kabinis will be able to play the new less demanding games from ea 2014. Thats a huge advantage for ea. The profit in the games is probably 10-30 times more than the apu. Content is king.
This is damn smart for ea.

 

[3DVagabond]

It is all speculation that they will code for PC first. No matter how great mantle is on PC, the vast majority of money will be made in consoles, so my best guess would be that they would still code for consoles first. At least that would be my expectation until developers clearly state otherwise.

The problem for developers is that they still have to code for 2 separate codepaths, and probably will sell very few more games for PC than if they only had to code for DX. What I mean is if mantle shows a good performance improvement, I doubt that will result in more PC versions being sold, just a shift of the market share of graphics cards, which developers dont care about, unless one of them is pumping them loads of cash.

Suppose Mantle reduces the hardware requirement significantly? Would that not increase the available market for the game? There's more facets to this than simply discrete market share % between the brands.