Why Mantle is godesend for people who stream using 1 PC

[Videophile]

Hey guys,

So BF4 recently got its mantle API engine update. So far, the drivers from AMD arent out yet, but so far it has shown that weaker CPU's have higher performance boosts in percentages than better CPU's like say a 4770k or a 4930k.

However, this got me thinking; I stream a lot of BF4, using one PC via game capture. Now BF4 is a CPU intensive game, and this update means I can devote more CPU resources to OBS since BF4 will run better, with lower CPU usage.

See what im getting at here? Of course, we have to leave it up to Jim to see if OBS can hook into the new renderer or not, and if not, just use monitor capture with windowed border-less(Thats if you are on Windows 8.1)

-Shrimp

 

[vbdkv]

Yeah I already have this installed, just waiting for the official Mantle drivers. Can't friggin wait.
http://store.steampowered.com/app/267130/

I tried the leaked ones and I must say, the difference is like night and day. Mantle performs so much better compared to DirectX, it's almost like it's a dream or something.

 

[Joppsta]

Or you can get an Nvidia graphics card with an inbuilt H.264 encoder and use shadowplay..

Mantle is awesome, don't get me wrong.. but I think I'll take the encoder. Assuming that Nvidia does the smart thing and makes it open so programs like OBS can utilize it. Practically no performance loss?

Yes please.

Cannot even begin to imagine what it would be like if Nvidia adopted Mantle themselves.. although I don't see them doing it out of pride and probably not wanting to rely on AMD to provide an API for people to use their products on.

 

[Krazy]

I'm pretty sure you guys are vastly overestimating the gains you will get with Mantle. It's not going to suddenly and drastically reduce CPU cycles, and it's much more meant to enhance performance of their APUs.

The other thing is I really don't see this getting wide adoption. It requires time and resources from developers to adjust their game for an API that doesn't even run on the majority of the market's hardware. I wouldn't be surprised if the only reason BF4 has it is because AMD paid a lot of money for it.

edit: and this thread belongs in the Feedback and Suggestions forum

 

[vbdkv]

I'm pretty sure you guys are vastly overestimating the gains you will get with Mantle. It's not going to suddenly and drastically reduce CPU cycles, and it's much more meant to enhance performance of their APUs.

The other thing is I really don't see this getting wide adoption. It requires time and resources from developers to adjust their game for an API that doesn't even run on the majority of the market's hardware. I wouldn't be surprised if the only reason BF4 has it is because AMD paid a lot of money for it.

edit: and this thread belongs in the Feedback and Suggestions forum

It will run on every major console, how is that not a market? ..
Anyway try the demo and see the difference for yourself. Good luck running all that without Mantle (a.k.a. without AMD hardware). Of course if you don't have AMD hardware to begin with, don't bother with the demo. It'll probably kill your pc anyway.

 

[Krazy]

Well, the topic in question is PC gaming, which is dominated by Intel and NVIDIA. It doesn't really matter what consoles do, but if developers want to port console titles to PC, it's a lot easier on them to do everything in DirectX from the start and have a unified platform.

Anyway, with AMD's track record of buggy drivers and problematic software, I'm not really holding my breath on anything game changing here.

 

[vbdkv]

Number of drivers that actually killed graphics cards:
Nvidia = 3
AMD = 0

Buggy drivers? Not likely :)
Also if something is being dominated by a certain brand, it's because they offer better performance. Mantle completely and utterly change that. I'm no fanboy at all, I'm in the nvidia and AMD camp, but I'm not utterly blind to the gigantic shift in gaming Mantle is going to provide. My link above is an early unfinished alpha benchmark, and it already blows DirectX completely out of the acquisition.

Good things are incoming for AMD that is certain. They also provide the APU for all major consoles in case you didn't know. The PC market is dictated by consoles these days, all we get are ports of AAA games.

But hey, if you wanna stick with intel and nvidia, go right ahead. I'm not judging anyone. I would have done the same a few years back when Matrox was dominating the market and 3dfx was just a phase ..

Ever heard of Matrox? ........... dot dot dot .. 3dfx rings a bell? .. There you go :)

 

[Lain]

If AMD makes the API proprietary then it will not be possible to capture it or use it for rendering. You will be forced to use directx or opengl on the games that can use it if so.

Whether it's actually more than hype has yet to be determined.

 

[vbdkv]

Without Mantle I was getting around 27FPS. With Mantle I was getting around 43FPS. That's a huge increase in performance for absolutely free. And its only an early alpha build of that engine without much optimization done and a beta driver, so it's definitely more than hype.

Check it out yourselves

1) Download beta driver with mantle enabled
http://support.amd.com/en-us/download/d ... ows+7+-+64
2) Install the swarm tech demo
http://store.steampowered.com/app/267130

 

[AndehX]

I smell a fanboy.

 

[Boildown]

It seems Mantle reduces CPU usage instead of GPU usage, which is certainly a surprise to me.

http://hardocp.com/article/2014/02/03/a ... tlefield_4
http://www.anandtech.com/show/7728/batt ... le-preview

But seriously, you can't capture it with OBS when Mantle is active? I didn't think that would be a problem.

 

[Lazze]

Mantle is not going to cause a "gigantic shift in gaming". It was primarily developed for increasing the performance of their APU's. Mantle is NOT magic, and it will not be widely adopted by AAA games.

Mantle might be the future, but for now it is insignificant. Developers have no reason to implement it, it's time and money down the drain for what is essentially a small FPS increase. Anyone saying that Mantle is the most important thing in gaming right now is a fanboy. Don't get me wrong, I think it's awesome technology and I would love to see it succeed, but right now it is not nearly as important as some of you make it out to be.

 

[Videophile]

Im just saying, for those people who want to stream and play BF4 at same time on the same PC, mantle will allow for more CPU usage to go towards OBS. (you can play in windowed fullscreen and do monitor capture)

Btw, OBS detects BF4 when running with the mantle renderer, but when added as a source, its a black screen.

-Shrimp

 

[Boildown]

I think the idea is that since both new consoles are running AMD processors, most of the work for supporting Mantle will already be done when porting those games over to PC. Add to that some major titles that will add support to their game development, including Star Citizen, which is the game I'm currently looking forward to. Personally I remember when Glide was hugely important, Mantle could become the next Glide.

Truth is, I actually agree with the OP in principle, now that I know what Mantle does. It frees up CPU cycles for those people who need it the most, those with lower-end CPUs. People who couldn't stream and play before will be able to, assuming all of Jim's concerns with the Mantle API are taken care of.

My background: I haven't run an AMD GPU since the Radeon Pro 9700... oh wait, that was ATI. I've never run an AMD GPU, been Nvidia the entire time since I finally retired that one ATI. Yet I can acknowledge that the early Mantle results look better than expected if you're a low to middle-end CPU'd OBS streamer, outside of the API problem.

 

[Lain]

If AMD releases it to the public, I will attempt to implement support for it.

However, let me state clearly for the record my disappointment with their decision to make yet another API. They could have simply done the same with OpenGL extensions, and any claim otherwise is pretty much a joke as vendors are allows to make whatever sort of extensions they want for their devices. nvidia already allows the same sort of features through OpenGL extensions. Don't be so quick to buy in to hype.

If they never release it to the public, then you can forget about ever being able to capture mantle games. You may as well turn it off if that ends up being the case.

 

[Lazze]

Personally I remember when Glide was hugely important, Mantle could become the next Glide.

There is a very good reason why Glide is not important anymore, it was proprietary to 3DFX graphics cards. The exact same thing is happening to Mantle. It's only for AMD cards, that limits the market significantly right there. If Nvidia does not implement Mantle, then I would consider it dead. Like Jim said, why have another API when you can do the same with the existing one.

 

[Boildown]

Personally I remember when Glide was hugely important, Mantle could become the next Glide.

There is a very good reason why Glide is not important anymore, it was proprietary to 3DFX graphics cards. The exact same thing is happening to Mantle. It's only for AMD cards, that limits the market significantly right there. If Nvidia does not implement Mantle, then I would consider it dead. Like Jim said, why have another API when you can do the same with the existing one.

I do not think that Mantle will be a deciding factor forever, even in the most AMD-optimistic scenario. But if they have a good two years of relevance, it will be a big deal. Calling Mantle dead before it even gets started is just silly. Glide was killed by Nvidia, when 3dfx sold out, by the way.

As for OpenGL, if these performance increases could be done with the existing API, why haven't they done it already? Indeed, why have another API if another way already existed? The best answer is that Mantle provides some capabilities that can't exist any other way. If you say, "because they are being mean to Nvidia", well then I say, "good for them". Then it should be simple for Nvidia to duplicate the results with OpenGL. More likely, is that it can't be done in OpenGL, or doing it in OpenGL gives away their advantage and research. And if Nvidia has to re-invent another Glide to compete with Mantle, because OpenGL and DirectX both aren't good enough, then that sucks that standards are failing, but perhaps they deserve to fail.

Is Nvidia sharing G-Sync? No, and I wouldn't expect them to either. AMD fanboys are crying that Nvidia isn't playing fair or something regarding G-Sync, and the opposite argument is being said about Mantle. Both camps are wrong to cry about this. Of course AMD should disclose their API, all this said, if they don't, then I'll see Mantle as being A Bad Thing, not until.

 

[Videophile]

Personally I remember when Glide was hugely important, Mantle could become the next Glide.

There is a very good reason why Glide is not important anymore, it was proprietary to 3DFX graphics cards. The exact same thing is happening to Mantle. It's only for AMD cards, that limits the market significantly right there. If Nvidia does not implement Mantle, then I would consider it dead. Like Jim said, why have another API when you can do the same with the existing one.

Mantle is not proprietary. They said it will run on any GCN card, but with some adaption could run on Nvidia GPU's as well. Im pretty sure thats where they are headed anyways.

 

[Lazze]

As for OpenGL, if these performance increases could be done with the existing API, why haven't they done it already? Indeed, why have another API if another way already existed? The best answer is that Mantle provides some capabilities that can't exist any other way. If you say, "because they are being mean to Nvidia", well then I say, "good for them". Then it should be simple for Nvidia to duplicate the results with OpenGL. More likely, is that it can't be done in OpenGL, or doing it in OpenGL gives away their advantage and research. And if Nvidia has to re-invent another Glide to compete with Mantle, because OpenGL and DirectX both aren't good enough, then that sucks that standards are failing, but perhaps they deserve to fail.

As Jim said, what AMD has done with Mantle can be implemented with OpenGL extensions. The reason Nvidia has not done something similar is that it still requires work from the developers to implement into their games. They can't simply "duplicate the results with OpenGL", that's not how it works. Either way, it still requires work from developers just like Mantle does. Nvidia could duplicate the results of Mantle with OpenGL, but it would add another level of complexity for developers, just like Mantle. I think it simply comes down to AMD being first with the idea, Nvidia could simply follow up if they wanted.

The issue with this is that we then have two companies doing the same thing with different API's. That's bad for the consumer AND the developers. I just don't see why Nvidia would implement Mantle rather than making something themselves. AMD made Freesync (their G-Sync alternative) even though Nvidia has said that they would try to make G-Sync available for all vendors, and that just shows what direction AMD is heading. I don't blame them, but it makes me doubt that Mantle will become available for everyone.

 

[Boildown]

See that's the thing. For those games being built for AMD graphics on either new console, and then ported to the PC, there isn't much additional effort for the gain. I'll see your point for PC-original games, if it really takes a lot of additional work.