AMD Mantle Interview with Oxide Games' Dan Baker

[Rasch]

Mantle gave nothing yet.BF4 runs flawlessly without it anway soo...Thief will have it with a dinosaur age engine.We will still buy high end gear regardless of the API and bottlenecks.

 

[SiDeards73]

Everytime i see one of these threads my faith in humanity dies a little more

 

[=XDC=FluphyBunny]

Mantle gave nothing yet.BF4 runs flawlessly without it anway soo...Thief will have it with a dinosaur age engine.We will still buy high end gear regardless of the API and bottlenecks.

Yes but your high end gear would still potentially gain from a fully working Mantle API in future games. Taking some of the processing load from the cpu would free it up to work on other things ie AI in single player games.

Also all this lock in/lock out business isn't the onus on Nvidia as to whether they use Mantle or not rather than down to AMD?

If Nvidia put their support behind Mantle it would be better for everyone as the devs would fall over themselves to support it.

 

[Marine-RX179]

Mantle gave nothing yet.BF4 runs flawlessly without it anway soo...Thief will have it with a dinosaur age engine.We will still buy high end gear regardless of the API and bottlenecks.

That's the thing- Thief is using unreal engine 3 right? Which it would probably mean is that the game won't use more than 3-4 CPU cores, so people with a FX or lesser CPU would bottleneck mid-high to high-end graphic card and might struggle to hold 60fps (not because of graphic card not fast enough); Mantle (if work properly) might solve that issue and reduce/remove that CPU bottleneck (and I haven't even mention people such as myself who game on 120/144Hz monitor and how CPU bottleneck can be a pain in the rear).

IMO Mantle would probably benefit old engine that uses few CPU cores more than new engine that would address and use say- 8 threads or more.

 

[Rasch]

We need 1-2years to really see what it worth,but good points Marine.

 

[Orangey]

Right?

So how is this any different? Because it will do in the future?

Maybe I've been around too long...I'm not doing this on purpose. I'd never praise PhysX as a technology, I'd rather it left to everyone else as much as anyone.

I dunno, I was highly skeptical they would/could ever support NV but apparently it's possible. Just waiting to see where it goes.

 

[tommybhoy]

I dunno, I was highly skeptical they would/could ever support NV but apparently it's possible. Just waiting to see where it goes.

AMD can't support Mantle on Nvidia hardware due to driver EULA's, Nvidia would have to adopt Mantle and develop support themselves.

Mantle support on Nvidia will be Nvidia's decision not AMD's.

 

[Silent_Scone]

IMO Mantle would probably benefit old engine that uses few CPU cores more than new engine that would address and use say- 8 threads or more.

Mantle's best bet is with games such as Star Swarm where by it's built from the ground up. Eventually there will be a game that has money thrown at it which will be built with Mantle very much in the forefront.

This is where it starts to benefit all scales of hardware from low end to high. At least those that are able to use it. I'd expect higher resolution textures being fed in, more draw distance, seemlessly flowing large scale combat such as that in the Oxide bench but with detailed human NPCs.

This is what I'd like to see out of what AMD have promised. What we're seeing at the moment is the initial baby steps towards this, but this is the downfall of all that is PC land. If you want to address DirectXs downfall, it's not entirely productive to remedy an industry level problem with one that will most likely cause a consumer level one. The odds on Nvidia adopting Mantle are very slim, and this is a key element to Mantles success. Apparently though I am a raging fanboy for pointing out what at least to me is the flaming obvious.

The good thing about OpenGL compared with both DirectX and Mantle, is that it's a big cooking pot for everyone to add their ten pence. Intel, AMD, NV and the like. This will likely never be the case with Mantle. We need a hippy loving programming language built from scratch and open to all from word go.

 

[Final8y]

The only thing worth reading in this thread was the Article.

 

[ICDP]

Good article, thanks for posting.

 

[Frenzy]

The only thing worth reading in this thread was the Article.

This, thanks for the article.

 

[Marine-RX179]

Mantle's best bet is with games such as Star Swarm where by it's built from the ground up. Eventually there will be a game that has money thrown at it which will be built with Mantle very much in the forefront.

This is where it starts to benefit all scales of hardware from low end to high. At least those that are able to use it. I'd expect higher resolution textures being fed in, more draw distance, seemlessly flowing large scale combat such as that in the Oxide bench but with detailed human NPCs.

This is what I'd like to see out of what AMD have promised. What we're seeing at the moment is the initial baby steps towards this, but this is the downfall of all that is PC land. If you want to address DirectXs downfall, it's not entirely productive to remedy an industry level problem with one that will most likely cause a consumer level one. The odds on Nvidia adopting Mantle are very slim, and this is a key element to Mantles success. Apparently though I am a raging fanboy for pointing out what at least to me is the flaming obvious.

The good thing about OpenGL compared with both DirectX and Mantle, is that it's a big cooking pot for everyone to add their ten pence. Intel, AMD, NV and the like. This will likely never be the case with Mantle. We need a hippy loving programming language built from scratch and open to all from word go.

Sorry, but I was specifically addressing the CPU limitation aspect and what Mantle can do to help in those area rather than talking about Mantle as we whole.

As for OpenGL, if it can really make that much of difference to help with boosting performance, I don't think anyone- AMD or Nvidia user would be against it. However based on what baker said regarding OpenGL, I'm not sure if it would actually provide that huge of an improvement over dx.

I don't know about you, but I myself definitely don't know nor would I pretend to know as much as or more than the developers do about coding for OpenGL. I think most people only have experienced using OpenGL in games or know about the concept/theory of how it works, rather than having first hand experience coding for it for a game. But yea if at all possible, bring on OpenGL that would make Mantle looks like a joke, and all of us consumers would be laughing till our jaws are stuck

 

[Silent_Scone]

Sorry, but I was specifically addressing the CPU limitation aspect and what Mantle can do to help in those area rather than talking about Mantle as we whole.

As for OpenGL, if it can really make that much of difference to help with boosting performance, I don't think anyone- AMD or Nvidia user would be against it. However based on what baker said regarding OpenGL, I'm not sure if it would actually provide that huge of an improvement over dx.

I don't know about you, but I myself definitely don't know nor would I pretend to know as much as or more than the developers do about coding for OpenGL. I think most people only have experienced using OpenGL in games or know about the concept/theory of how it works, rather than having first hand experience coding for it for a game. But yea if at all possible, bring on OpenGL that would make Mantle looks like a joke, and all of us consumers would be laughing till our jaws are stuck

I didn't praise OpenGL as an API, just as an open standard. We'll see in time though...

 

[Ragnahex]

Thanks Greg for the article it's an interesting read

 

[Paul_cz]

I pray to all PC Gods that Mantle will take off and games will support it. I want devs to use all power in my 280X, not just a part of it, dammit.

 

[Geck0]

Mantle's best bet is with games such as Star Swarm where by it's built from the ground up. Eventually there will be a game that has money thrown at it which will be built with Mantle very much in the forefront.

You just proved that you didn't bother reading the article. Oxide stated that they spent a huge amount of time optimising DX and very little on Mantle, even so Mantle still dwarfs DX in performance.

There has not yet been a game built around Mantle "from the ground up".

 

[Silent_Scone]

Nope. Star Swarm has been built with Mantle in mind from day one. He's saying they've spent a lot of time optimising for DX also if not more, that's all.

Anything else, Lassy?

 

[Steampunk]

Actually, Oxide staff have said on several occasions that Nitrous was built from the ground up for DX, is highly optimised for DX, but Mantle code was added in two man-months with little optimisation late in the day.

The only thing that was built in "from day one" is the idea that Nitrous would be multi-thread for modern hardware and for strategy/RTS style games.

 

[Geck0]

Nope. Star Swarm has been built with Mantle in mind from day one. He's saying they've spent a lot of time optimising for DX also if not more, that's all.

Anything else, Lassy?

Baker: This depends on how exploitative you are, and the specifics of your engine. For us, we have been completely limited in what we could do by driver overhead problems. We were actually making decisions where we traded GPU performance for CPU–that is, we’d end up doing things that are slower on the GPU, because we could get away with less driver overhead. When you talk about building an FPS, you probably spend much of your time optimizing for the GPU; when you try to build an RTS, you end up optimizing for the driver overhead. Nitrous is a new breed of rendering system. Oxide’s specialty is high throughput.

When you look at Star Swarm, it's really a testament to brute force. For us, we can see cases where Mantle is many times faster, with especially big differences as we add more cores and slow down their clock speeds. We wouldn't expect most games to necessarily see this, as it will happen in cases where you have a really efficient, high throughput engine, but it will certainly make an impact everywhere. We aren't set up to do very precise testing, so we'd rather others do the analysis on this. However, we'd like to point out that our Direct3D performance is absolutely outstanding, relative to what is expected. We have spent a huge amount of time optimizing around D3D, and we feel we are actually pretty biased in D3D’s favor. Mantle, on the other hand, we've spent far less time with and currently have only pretty basic optimizations. But Mantle is such an elegant API that it still dwarfs our D3D performance.

DERP

 

[GoogalyMoogaly]

Yes but your high end gear would still potentially gain from a fully working Mantle API in future games. Taking some of the processing load from the cpu would free it up to work on other things ie AI in single player games.

Also all this lock in/lock out business isn't the onus on Nvidia as to whether they use Mantle or not rather than down to AMD?

If Nvidia put their support behind Mantle it would be better for everyone as the devs would fall over themselves to support it.

I don't see it. I think in the long run it'd be bad for Nvidia. Giving that much control to your main rival just seems silly. Sort of like entering into a quiz against someone and then letting them pick the questions. Then if you did well on those questions, they could change the questions.

I know it's been said that Mantle will be supported on non-GCN cards, but has it been said that they will definitely see the same performance benefit?
What if AMD's non-GCN cards and Nvidia cards see a 10% boost under Mantle while GCN cards see a 40% boost under Mantle?
Is it really a good idea for Nvidia to support Mantle ensuring AMD have a considerable performance gap advantage on all new cards (which is what they're both trying to sell)? Or might it be better for Nvidia to avoid Mantle and support their own API or libraries for existing API to get the most performance they can out of them and hope the fact they don't adopt it Mantle never really gathers the support AMD want?