The AMD Mantle Thread

[krumme]

I would love to hear your definition of objective "data" because clearly actual website reviewed performance numbers do not exist. Was BF4 mantled reviewed by AT? No, I don't think so. All we have so far are marketers and parties paid for by an AMD GE contract talking about Mantle. That isn't necessarily a bad thing, but marketers and paid parties will not state objective facts - their interest is in making Mantle sound like the best thing ever. Which it might be, but we don't know.

I'll wait and see - I for one think AMD made the right move in creating mantle, but by the same token they will be utterly retarded to make it open and available for all vendors. Anyway, back to the point. There is no "data". Just marketing. It could be the best thing ever or it could be meh. Until websites review it we don't know.

Objective data is the best measurable important data you can get where you are at a given time in a process. Waiting until the end to search for "objective" data will kill your business before its started.

What we can see now when mantle is just starting:

2 engines
20 games
One game intented only for PC !
Oxide saying way above 100.000 batch - a factor of about 40 times more (easy to measure but they could be lying)
2 man months - (again they could be lying)

Have you seen the Q&A from apu13? Do you think its marketing answers? - or do you think they are lying?

 

[blackened23]

They're being paid by contract under AMD GE. So are they going to say anything bad? What do you think they're going to say? "Hey so AMD paid us a hefty contract but we think Mantle is terrible". No, they're being paid so they're going to say good things. So you're basically acknowledging that there isn't any "data". Like I said, we'll see what Mantle brings when it is in use which is December at the earliest for BF4. It could be great or it might not - I think AMD made the right move in creating Mantle. But there is no actual objectively measured performance data. Only marketers and paid parties (under AMD GE) trying to trump it up.

 

[krumme]

Since this is a thread about Mantle it seems the proper place to speculate about Mantle.
I haven't seen many threads recommending AMD GPUs over NVIDIA GPUs based on Mantle.
If people were doing so that would be the proper place to demand hard data or shut up.

I understand that and sometimes i agree. But i dont think its always that simple.

Take a person asking for advice for upgrading to a 770 from an 480. They intend to play a lot of bf4 at 1080p with their 3570, and ask if 2 or 4 gb is the right choice. What is the good response?

For many games mantle can be a game changer, because it not only helps gpu but also can have serious impact on cpu perf (eg. could be 64 multiplayer bf4, rts like sc). Pretending it does not exist until the game is there - eg SC - is making a big mistake. A mistake you wouldnt do yourself. Ofcourse you adapt to predictions and asumptions. We always do. Its double standards not to mention mantle when its pretty much there. Especially now when there is no cost associated in chosing it.

 

[krumme]

They're being paid by contract under AMD GE. So are they going to say anything bad? What do you think they're going to say? "Hey so AMD paid us a hefty contract but we think Mantle is terrible". No, they're being paid so they're going to say good things. So you're basically acknowledging that there isn't any "data". Like I said, we'll see what Mantle brings when it is in use which is December at the earliest for BF4. It could be great or it might not - I think AMD made the right move in creating Mantle. But there is no actual objectively measured performance data. Only marketers and paid parties (under AMD GE) trying to trump it up.

Go and see the Q&A for yourself instead of all those asumptions and strawmen. You might fint something new. Its not the usual ppt marketing bs and that makes it actually worth watching. Its by far the best information we have about the devs. motives as there is plenty of examples also.

I presented you with plenty facts. 2 engines 20 games. No 2014 performance engine DX only until now. lol. Thats also a fact.

 

[GaiaHunter]

They're being paid by contract under AMD GE. So are they going to say anything bad? What do you think they're going to say? "Hey so AMD paid us a hefty contract but we think Mantle is terrible". No, they're being paid so they're going to say good things. So you're basically acknowledging that there isn't any "data". Like I said, we'll see what Mantle brings when it is in use which is December at the earliest for BF4. It could be great or it might not - I think AMD made the right move in creating Mantle. But there is no actual objectively measured performance data. Only marketers and paid parties (under AMD GE) trying to trump it up.

Don't AMD get games to add for their bundles as well?

 

[VulgarDisplay]

They're being paid by contract under AMD GE. So are they going to say anything bad? What do you think they're going to say? "Hey so AMD paid us a hefty contract but we think Mantle is terrible". No, they're being paid so they're going to say good things. So you're basically acknowledging that there isn't any "data". Like I said, we'll see what Mantle brings when it is in use which is December at the earliest for BF4. It could be great or it might not - I think AMD made the right move in creating Mantle. But there is no actual objectively measured performance data. Only marketers and paid parties (under AMD GE) trying to trump it up.

If this was the case repi would not be showing his face at nvidia events.

 

[PPB]

I perfectly understanded what you wrote ub the first post, i just dont agree.

You might want to consider the possibility i see something you dont.

Give me an example where the above situation is happening 2014. Give me an example of a game.

And then understand Mantle as AMD have said it:

If you dont care about performance. Mantle is not for you.

If you dont care about cross platform compatability. Mantle is not for you.

Actually there will be plenty games, eg online, that falls in the above category. Everything will continue as usual here. Nothing have changed.

But Mantle is not targeting those games.

Mantle is all about performance

The flexibity and functionality in mantle is all for performance.

Its like talking to a brick walk. Unless brick walls dont come back with non sequiturs that have little to do with your point.

Ehm, im off, have fun responding nonsense to the rest that still care.

 

[krumme]

Its like talking to a brick walk. Unless brick walls dont come back with non sequiturs that have little to do with your point.

Ehm, im off, have fun responding nonsense to the rest that still care.

Yeaa. Then while you are off good luck to find a game where the saved development cost using mantle is getting used at charity dx development and not goes into the devs pockets.

 

[krumme]

What is actually interesting with mantle from a gamers peespective is the possibility of new games - not the performance.

Obviously the small batch problem can give somethink but i think the extended control mantle gives the devs can give us somethink that is not expected comming in 2015 - 2017.

That is the best perspective especially for gamers like us. We have usually plenty of power under the hood to get the full experience from the games.

 

[MeldarthX]

What is actually interesting with mantle from a gamers peespective is the possibility of new games - not the performance.

Obviously the small batch problem can give somethink but i think the extended control mantle gives the devs can give us somethink that is not expected comming in 2015 - 2017.

That is the best perspective especially for gamers like us. We have usually plenty of power under the hood to get the full experience from the games.

You're going to want to add another game engine - which will translate into more games to that list

 

[Borealis7]

98 pages of worthless squabbling and not a single performance number! come on i'm tired of waiting!

 

[Noctifer616]

98 pages of worthless squabbling and not a single performance number! come on i'm tired of waiting!

Late December is the date for the Bf4 patch, so at least another 3 weeks I guess.

 

[Paul98]

98 pages of worthless squabbling and not a single performance number! come on i'm tired of waiting!

They do give performance numbers such as the reduced draw call overhead.

But the reason you aren't getting an overall performance picture is that it's a new API currently in alpha. The only for sure performance gains are from the removal of the DX overhead. But that is only a small part of what can be done with mantle. So it's not that mantle will just give around a certain gain in all games. It's going to depend on the game and what they are able to take advantage of from mantle. You could have one game which is 150% faster and another only 15%.

 

[MeldarthX]

Paul98 I don't think we'll ever see as low as 15% simply because of being able to bypass DX and access the hardware through mantle a lot easier.

Usual 30% min is I think what we will see - we got some pretty good details from Q&A and the new game engine demo; that's just the tip of the iceberg.

 

[Paul98]

Paul98 I don't think we'll ever see as low as 15% simply because of being able to bypass DX and access the hardware through mantle a lot easier.

Usual 30% min is I think what we will see - we got some pretty good details from Q&A and the new game engine demo; that's just the tip of the iceberg.

It all depends on the game and how it's used, I would agree that the average is going to be much higher though.

 

[psoomah]

With the GPU13 and APU13 DICE presentations Johan Andersson is telling developers and publishers their cherished Holy Grail of write once, cheap and easy high quality close to the metal ports to all platforms is doable and achievable IF enough get on board with Mantle and ACT to make that happen. What could be a more compelling common goal for developers and publishers to unite behind and act in concert to make happen?

Microsoft might remain an outlier leveraging it's Xbox One DX based rendering path into it's closed Xbox/PC store ecosystem, hence the APU13 DICE slide that contained this line: "Mantle and PlayStation 4 will drive our future Frostbite designs & optimizations" to emphasize Microsoft does not have to be on board for Mantle to succeed. Might because three or four years out, if Mantle is on track to become the industry standard programming model and Microsoft remains the sole remaining outlier to a unified programming model, programming a DX path for the Xbox One will be considered an unnecessary added expense to developers and publishers. Microsoft is going to come under intense pressure to get on board with Mantle.

EA of course is looking at the future possibilities with $$$ signs $$$ in their eyes, especially after the profits eating grind of porting between the PS3, Xbox 360 and PC.

Johan Andersson has stated Mantle is extensible to, and he WANTS it on AMD, Intel, Nvidia, ARM, Windows, Linux (Steam OS), Apple and Android. DICE made a point of including these lines in their APU13 slide presentation:
......

"Mantle + Steam OS = powerful combination."

"Major opportunity with Mantle - leap frog GL4, DX11
- For mobile SoC vendors
- For Goggle and Apple"
......

It also emerged Mantle was built to accommodate future architectures. Once implemented that HOLY GRAIL is good into the foreseeable future.

To top it off Mantle coming late to BF4 means it can be retrofitted to games initially released without it. Publishers and developers coming late to Mantle can include it on ALL their next gen games, even those already released.

Mantle is literally a doable and achievable Holy Grail Swiss Army Knife for developers and publishers.

Mantle is Johan's baby, he went to AMD with the vision and with the console wins as their leverage, he got them on board. He is the obvious evangelist here and is doubtless spreading that Holy Grail potential and promise message to the other developers and publishers. The more developers and publishers that get on board and push Mantle the more pressure there will be for Nvidia and Intel to get on board. The critical question will be if and when AMD makes Mantle an open standard and available to it's hardware competition. I see that timing as a balancing act for AMD ... stretch out that release as long as possible while still ensuring it becomes the industry programming standard. I would guess AMD getting a Beta GCN SDK out by March/April and by mid year 2014 announcing a 1Q 2015 release to open standard date so it has the 2014 holiday Mantle game releases to itself.

Anyhoo, just to point out a critical and extremely compelling reason why the big developers and publishers in particular (once they fully understand the potential) would have every incentive to get on board with Mantle and put it in ALL their games to apply maximum pressure to Nvidia and Intel to back it when it is made available ... and, say, three years out if Mantle has released as open standard and Nvidia or Intel still resist getting on board, those developers and publishers will have zero incentive to optimize for those Intel and or Nvidia or provide timely builds for their drivers.

Why help those impeding reaching that $$$ Holy Grail $$$?

 

[krumme]

You're going to want to add another game engine - which will translate into more games to that list

Lol yeaa you are right. Actually when you write it i think we will have to wait for the 2015 - 2017 engines meaning games at least 2 years from now, until we see the creative potential mantle could unleash?
Probably 4 to 6 years. There is great potential in giving the devs the control instead of using their time on idiotic debugging and eg guessing how much ram a texture uses

 

[blastingcap]

With the GPU13 and APU13 DICE presentations Johan Andersson is telling developers and publishers their cherished Holy Grail of write once, cheap and easy high quality close to the metal ports to all platforms is doable and achievable IF enough get on board with Mantle and ACT to make that happen. What could be a more compelling common goal for developers and publishers to unite behind and act in concert to make happen?

Microsoft might remain an outlier leveraging it's Xbox One DX based rendering path into it's closed Xbox/PC store ecosystem, hence the APU13 DICE slide that contained this line: "Mantle and PlayStation 4 will drive our future Frostbite designs & optimizations" to emphasize Microsoft does not have to be on board for Mantle to succeed. Might because three or four years out, if Mantle is on track to become the industry standard programming model and Microsoft remains the sole remaining outlier to a unified programming model, programming a DX path for the Xbox One will be considered an unnecessary added expense to developers and publishers. Microsoft is going to come under intense pressure to get on board with Mantle.

EA of course is looking at the future possibilities with $$$ signs $$$ in their eyes, especially after the profits eating grind of porting between the PS3, Xbox 360 and PC.

Johan Andersson has stated Mantle is extensible to, and he WANTS it on AMD, Intel, Nvidia, ARM, Windows, Linux (Steam OS), Apple and Android. DICE made a point of including these lines in their APU13 slide presentation:
......

"Mantle + Steam OS = powerful combination."

"Major opportunity with Mantle - leap frog GL4, DX11
- For mobile SoC vendors
- For Goggle and Apple"
......

It also emerged Mantle was built to accommodate future architectures. Once implemented that HOLY GRAIL is good into the foreseeable future.

To top it off Mantle coming late to BF4 means it can be retrofitted to games initially released without it. Publishers and developers coming late to Mantle can include it on ALL their next gen games, even those already released.

Mantle is literally a doable and achievable Holy Grail Swiss Army Knife for developers and publishers.

Mantle is Johan's baby, he went to AMD with the vision and with the console wins as their leverage, he got them on board. He is the obvious evangelist here and is doubtless spreading that Holy Grail potential and promise message to the other developers and publishers. The more developers and publishers that get on board and push Mantle the more pressure there will be for Nvidia and Intel to get on board. The critical question will be if and when AMD makes Mantle an open standard and available to it's hardware competition. I see that timing as a balancing act for AMD ... stretch out that release as long as possible while still ensuring it becomes the industry programming standard. I would guess AMD getting a Beta GCN SDK out by March/April and by mid year 2014 announcing a 1Q 2015 release to open standard date so it has the 2014 holiday Mantle game releases to itself.

Anyhoo, just to point out a critical and extremely compelling reason why the big developers and publishers in particular (once they fully understand the potential) would have every incentive to get on board with Mantle and put it in ALL their games to apply maximum pressure to Nvidia and Intel to back it when it is made available ... and, say, three years out if Mantle has released as open standard and Nvidia or Intel still resist getting on board, those developers and publishers will have zero incentive to optimize for those Intel and or Nvidia or provide timely builds for their drivers.

Why help those impeding reaching that $$$ Holy Grail $$$?

Holy Rainbow Brite, Batman!

Anyway, as much as I hate EA, I'm glad that some companies are making more multi-core friendly engines and pushing Mantle. Even if Mantle fails and gets co-opted, its ripple effects will probably be good for the industry as a whole.

 

[blackened23]

Mantle is Johan's baby, he went to AMD with the vision and with the console wins as their leverage, he got them on board.

I'm under the impression that he was paid 8 million dollars by AMD and then got on board. What order that happened in, nobody knows, but the 8 million? Yeah i'd be on board too! Whatever you say AMD! Mantle, done!

On a serious note, if this helps BF4 i'll be the first to say well done. I just really want to see real results instead of power points. It is about time AMD attempted to create value adds like nvidia has done for years to their products..

 

[blastingcap]

I'm under the impression that he was paid 8 million dollars by AMD and then got on board. What order that happened in, nobody knows, but the 8 million? Yeah i'd be on board too! Whatever you say AMD! Mantle, done!

On a serious note, if this helps BF4 i'll be the first to say well done. I just really want to see real results instead of power points. It is about time AMD attempted to create value adds like nvidia has done for years to their products..

$8 MM was in part for all those BF4 keys. And others like Oxide and RobertsSpace have jumped on board saying Mantle is something they've been seeking for a long time. So no, this is not just AMD pushing Mantle, there are definitely some devs who are pulling, too. I'm also pretty sure Valve is drooling over the prospect of widespread Mantle support or something like it.

 

[psoomah]

Holy Rainbow Brite, Batman!

Anyway, as much as I hate EA, I'm glad that some companies are making more multi-core friendly engines and pushing Mantle. Even if Mantle fails and gets co-opted, its ripple effects will probably be good for the industry as a whole.

I hear you, but co-opted by whom?

Mantle is the complete package checking every conceivable Holy Grail box. AMD is not only the single vendor able to do this, but the only vendor that can even come close to doing so.

Microsoft is committed to it's DX path.
Intel has no presence in the AAA graphics space.
Nvidia has no x86 capability.

From a developer and publisher point of view there simply is no acceptable alternative to what Mantle brings to the table. Why settle for table scraps when AMD is bringing a perfectly cooked full on Turkey dinner extravaganza to the table?

 

[96Firebird]

Whoa... This is fun to read. :thumbsup:

 

[SiliconWars]

I'm under the impression that he was paid 8 million dollars by AMD and then got on board. What order that happened in, nobody knows, but the 8 million? Yeah i'd be on board too! Whatever you say AMD! Mantle, done!

Repi actually said that they (AMD) wouldn't have needed to pay anything for Mantle as they (the devs) asked for it.

Also, judging by the low ~2 man month development time, we'd be talking thousands, not millions. Barely even 10's of thousand dollars in developer resources I'd bet.

 

[psoomah]

I'm under the impression that he was paid 8 million dollars by AMD and then got on board. What order that happened in, nobody knows, but the 8 million? Yeah i'd be on board too! Whatever you say AMD! Mantle, done!

On a serious note, if this helps BF4 i'll be the first to say well done. I just really want to see real results instead of power points. It is about time AMD attempted to create value adds like nvidia has done for years to their products..

Both Johan Andersson and AMD are on record saying Andersson approached AMD with the vision that became Mantle.

 

[blastingcap]

Microsoft is committed to it's DX path.
Intel has no presence in the AAA graphics space.
Nvidia has no x86 capability.

Read definition 4: http://www.thefreedictionary.com/co-opted

20% is nothing to sneeze at, but Microsoft won't take this lying down. Direct3D and OpenGL if improved enough may narrow the gap enough that Mantle becomes irrelevant. If Mantle is say, 5% faster, it won't even matter. Even 10% is shaky. That's like going from 30 fps to 33 fps. 15% maaaybe. Once we're in that 20+% zone things get more interesting but it's not necessarily a game changer even at 20%.

x86 is irrelevant to this discussion, NV can come up with an answer to Mantle, though the problems is that they aren't in any consoles.. but they are banking on mobile and cloud gaming anyway and it is theoretically possible that being in mobile matters more than being in consoles, in the long run (AMD has no mobile presence). I do think NV made a huge error in not being in next-gen consoles, though. That may come to haunt them in more ways than just Mantle.

Intel will have a greater mobile presence as well even if they aren't in consoles. But I am not sure Intel would actively oppose Mantle in the first place, since anything that boosts PC gaming relative to consoles can only help them, seeing as how Intel is not in consoles.