The AMD Mantle Thread

[frozentundra123456]

Yes, but it should make better use of the GPU. No?

Shouldn't it depend on the game? If it lowers CPU load, I would think it should show more improvement in CPU limited games and relatively less where you are already limited by the gpu.

 

[Kenmitch]

Shouldn't it depend on the game? If it lowers CPU load, I would think it should show more improvement in CPU limited games and relatively less where you are already limited by the gpu.

Once the BF4 Mantle patch is released we'll get a good indication of performance gains in both scenarios I'd think.

 

[Black Octagon]

A week late to the party but clearly Mantle is going to make some new games far better performing and hopefully all of them. Now I can believe the 290x could be twice as fast as Titan in BF4.

Me too, but something is annoying me about this "twice as fast as Titan" claim. How can they really know? Especially when the game is still in beta form, both Catalyst and Forceware drivers for this game are immature, and the Mantle functionality is still to be developed?

Even if all goes as well as AMD is hoping, the final comparison to Titan's BF4 performance will still depend big time on how good a job NVIDIA's driver development team does, right?

 

[cytg111]

Me too, but something is annoying me about this "twice as fast as Titan" claim. How can they really know? Especially when the game is still in beta form, both Catalyst and Forceware drivers for this game are immature, and the Mantle functionality is still to be developed?

Even if all goes as well as AMD is hoping, the final comparison to Titan's BF4 performance will still depend big time on how good a job NVIDIA's driver development team does, right?

"twice as fast as Titan" .. That is good for bragging rights, but I dont see a reason for it happening. Mantle is something that lets AAA console ports run on cheap AMD APU's on the PC, nothing else.
Riddle me this, would a Mantle patch yield the same performance benefit on an APU and a system with a descrete Radeon (GCN/UMA whatever) card?
Coding FOR having the GPU *just next doors* is just not porting well to -> lets leave the cpu, get on a bus, adress the nice lad in pciexpress lane 1 'can i please come in', come in issue your requests and GTFO.
Latency latency latency ..
And if consoles is calling the shots.. then Im sorry to say, we get shi* on the PC front for the next decade also.

 

[MightyMalus]

I'm guessing that the improvement could/should be about the penalty you get from using DX or OGL instead of the actual metal, plus, the less amount of draw calls.

I'd say 10-20% by just using Mantle instead of DX/OGL in the same GCN GPU. Max, 50%. IMHO, I just don't think DX and OGL have that much overhead...but then again, considering the old complaints, it could be a significant difference indeed.

I still think this is the most genius thing that's been done in PC Gaming for years tho!

 

[ShintaiDK]

So considering that the GPU is entirely unrelated to any API. How can GPU performance increase with Mantle?

Game->DirectX API, AMD Driver->GCN
Game->Mantle API, AMD Driver->GCN

Game, API and Driver is CPU only.

 

[Skurge]

So considering that the GPU is entirely unrelated to any API. How can GPU performance increase with Mantle?

Game->DirectX API, AMD Driver->GCN
Game->Mantle API, AMD Driver->GCN

Game, API and Driver is CPU only.

You refuse to believe even though devs are confirming the will be a performance increase. How much of an increase is unknown at this point.

 

[Paul98]

So considering that the GPU is entirely unrelated to any API. How can GPU performance increase with Mantle?

Game->DirectX API, AMD Driver->GCN
Game->Mantle API, AMD Driver->GCN

Game, API and Driver is CPU only.

What are you trying to say here? That the API won't effect the real world FPS? Or that it won't increase the theoretical performance?

 

[smackababy]

So considering that the GPU is entirely unrelated to any API. How can GPU performance increase with Mantle?

Game->DirectX API, AMD Driver->GCN
Game->Mantle API, AMD Driver->GCN

Game, API and Driver is CPU only.

That isn't how it works though.

It is Game > DirectX High level API > DX Driver > AMD Driver > GCN
Mantle looks like it it is Game > developer API > Mantle Driver > AMD Driver > GNC

The potential for it to have performance is there, because the DX stack is rather large and can't be skipped. By allowing developers to do their own high level API for image processing before Mantle makes the draw calls, CAN give performance increase. However, that is almost entirely dependent on the developer's ability; which means 90% of those involved in game development can't take advantage of it.

 

[ShintaiDK]

What are you trying to say here? That the API won't effect the real world FPS? Or that it won't increase the theoretical performance?

Just saying that if anything, it only affects CPU.

 

[Rvenger]

What happens if it gives over a 30% increase? Will BF4 become an irrelevant game all the sudden for the naysayers?

 

[smackababy]

What happens if it gives over a 30% increase? Will BF4 become an irrelevant game all the sudden for the naysayers?

If it does give a 30% increase in certain games, then good for those with AMD cards. But, if the performance in DX is crippled just to highlight AMD, that is complete BS and awful for gamers.

 

[3DVagabond]

If it does give a 30% increase in certain games, then good for those with AMD cards. But, if the performance in DX is crippled just to highlight AMD, that is complete BS and awful for gamers.

Do you mean if the Devs cripple the games Dx performance to artificially inflate the performance of Mantle? If that's what you are worried about, I think you are concerning yourself about nothing. Although, Devs quite often do poor Dx ports already. I don't see how Mantle will change that, and if it is easier to do a quality port in Mantle than Dx, it's a good thing IMO.

 

[blackened23]

If it does give a 30% increase in certain games, then good for those with AMD cards. But, if the performance in DX is crippled just to highlight AMD, that is complete BS and awful for gamers.

This wont' happen because AMD is targeting Mantle towards multi platform game engines, they aren't targeting specific developers. There are like, 3 engines powering 85% of all multiplatform console and PC games - All of the engines in question will still retain high quality DX11 platforms. Dice isn't going to gut their DX11 implementation in Frostbite - they want NV users to play their game as well. I don't see this as being much different than Epic implementing full phsyx functionality into Ue3 and UE4 - just because they did that does not detract from their DX implementation used. Epic wants their DX implementation to be great, and generally speaking, it is. They don't want only 35% of their playerbase buying games, they want 100% to buy their games - therefore they will not ignore DX11.

 

[Paul98]

If it does give a 30% increase in certain games, then good for those with AMD cards. But, if the performance in DX is crippled just to highlight AMD, that is complete BS and awful for gamers.

I don't seem them planning to cripple the performance, rather that the porting over to DirectX as is happening now will cripple the performance. Compared to if it was specifically made for DirectX.

Where I see AMD coming into a problem is when they want to release new graphics cards. Are they going to be compatible or will they go back to DirectX? If there are good performance gains, it is almost going to be like a step forwards then one back. What I could see is that compatibility is kept, if there are new things added, just add more functionality for the PC market.

 

[Paul98]

Just saying that if anything, it only affects CPU.

Why do you say that, we already know that you get more access to the graphics card. Such as direct access to video memory.

 

[Shivansps]

This wont' happen because AMD is targeting Mantle towards multi platform game engines, they aren't targeting specific developers. There are like, 3 engines powering 85% of all multiplatform console and PC games - All of the engines in question will still retain high quality DX11 platforms.

How you even know that?

To Mantle to be taken seriusly it must be a game changer, if you still need to target DX whats the point? All the gains on the developer side just vanish and Mantle or DX becomes a nuisance. "Because of consoles" does not really work when developing a PC port.

Not to mention if you still targeting DX you will be never able to take full advantage of Mantle, because you must ensure the game will be fully playable on DX.

Add SteamOS and OpenGL intro the equation and things just gets worse.

 

[monstercameron]

How you even know that?

To Mantle to be taken seriusly it must be a game changer, if you still need to target DX whats the point? All the gains on the developer side just vanish and Mantle or DX becomes a nuisance. "Because of consoles" does not really work when developing a PC port.

Not to mention if you still targeting DX you will be never able to take full advantage of Mantle, because you must ensure the game will be fully playable on DX.

Add SteamOS and OpenGL intro the equation and things just gets worse.

no offense but if dice can pull it off they crytek and unreal can do it haswell
and then that is a majority of pc games and console ports...

 

[Abwx]

The question is how much perfs improvements would
made such a tool worthy of the spilled money...

30% is good but is it enough considering that it s
about what is brought by a new GFX generation
for all price points ?

Or simply by getting rid of your 150$ card
for a 200-220$ one..?.

 

[Shivansps]

Performance increase only does not matter for devs, DX9 should be long dead if so, we also whould be using more Linux and OpenGL as well.

no offense but if dice can pull it off they crytek and unreal can do it haswell
and then that is a majority of pc games and console ports...

Not gona happen as long you still need to target DX/OpenGL alongside it, and "mayority"? nope, consoles does not count when doing a pc port, PC hardware and market shares count, with that in mind you gona still be targeting mainly DX and/or OpenGL, at that point Mantle is just a "extra weight" to carry along. I dont see anyone using it unless AMD pay them, its even worse than Physx.

If we are talking about a Mantle only port things will be diferent, but that not gona happen unless GCN gets +75% of PC market share. Again consoles does not count when you target PC.
Thats is the part where the AMD argument of "makes things easier for devs" fells apart, even if you game is console only, and it runs on Mantle on console(still not clear about it), if you gona do a PC port, you have 2 options, do it Mantle for 15% of the market at best or do it DX for 100% of the market, if you do both, them Mantle is not advantage for devs.

You even stopped for a moment to wonder how less buggy could BF4 be if they have not wasted time in mantle at all? Do you know development times? We are not even sure if they need mantle for console version at all...

 

[3DVagabond]

Performance increase does not matter for devs, DX9 should be long dead if so, we also whould be using more Linux and OpenGL as well.

Consoles used Dx9. Dx9 will die now.

Not gona happen as long you still need to target DX/OpenGL alongside it, and "mayority"? nope, consoles does not count when doing a pc port, PC hardware and market shares count, with that in mind you gona still be targeting mainly DX and/or OpenGL, at that point Mantle is just a "extra weight" to carry along. I dont see anyone using it unless AMD pay them, its even worse than Physx.

If we are talking about a Mantle only port things will be diferent, but that not gona happen unless GCN gets +75% of market share.

You even stopped for a moment to wonder how less buggy could BF4 be if they have not wasted time in mantle at all? Do you know development times?

Do you know how much work it is to use Mantle on PC? You are making a lot of statements against Mantle. How much knowledge do you have of Mantle to back your position?

 

[VulgarDisplay]

I'm kind of curious how Mantle will effect latency in games as well. If many of the draw calls never have to waste time going to the CPU and just happen on the GPU I would think that the time it takes to render frames will be reduced. That could have a bigger impact on perceived performance than higher framerates would if you ask me.

It could lead to less input lag which would be nice for hardcore PC gamers.

 

[Shivansps]

Consoles used Dx9. Dx9 will die now.





Do you know how much work it is to use Mantle on PC? You are making a lot of statements against Mantle. How much knowledge do you have of Mantle to back your position?

Im software engineer, well not yet. But you can guess i have a good idea on how much time and cost takes to implement a new api, Mantle or OpenGL or anything else does not matter, still i can assure you than implement Mantle will be harder than implement OpenGL, its a new api, that needs training for devs, and if it has low level elements on it thats worse.
And if your game engine was not developed with mantle in mind it will be even worse.


Im not saying Mantle will not be great, but there is a lot of issues with it to implement in a real world, its not so easy as AMD makes you to belive.

 

[VulgarDisplay]

Im software engineer, well not yet. But you can guess i have a good idea on how much time and cost takes to implement a new api, Mantle or OpenGL or anything else does not matter, still i can assure you than implement Mantle will be harder than implement OpenGL, its a new api, that needs training for devs, and if it has low level elements on it thats worse.
And if your game engine was not developed with mantle in mind it will be even worse.


Im not saying Mantle will not be great, but there is a lot of issues with it to implement in a real world, its not so easy as AMD makes you to belive.

Ever heard of these?
http://en.wikipedia.org/wiki/Video_game_console

The majority of all developers and engine licensing companies are going to be optimizing to the metal of GCN already. Yes it will involve some learning, but that's true of all console generations. The difference is, all the learning they do for the xbone and ps4 where probably 80% of their sales are going to come from now directly applies to AMD hardware on the PC.

 

[Shivansps]

Ever heard of these?
http://en.wikipedia.org/wiki/Video_game_console

The majority of all developers and engine licensing companies are going to be optimizing to the metal of GCN already. Yes it will involve some learning, but that's true of all console generations. The difference is, all the learning they do for the xbone and ps4 where probably 80% of their sales are going to come from now directly applies to AMD hardware on the PC.

Ive already explained, it does not work that way, you may stop for one moment and start thinking about it, or not, not my problem.