• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Activision Technical Director - "We'll have to support MANTLE now"

artist

Banned
a94ut2.png

bw2uq2.png


https://twitter.com/wadetb/status/383266277818507264
 

artist

Banned
EA and Activision - that's two big players right there. AMD also has a foot in the door with Ubi, I'm sure they are making headway there as well. It'll be interesting to see the sort of pubs, devs AMD will get lined up for their Dev summit in Nov.
 

kortez320

Member
Having it as opengl extension would be much better. It's got more chance of surviving that way.

Yeah. This is kind of my problem with the whole thing.

I'm sure we will get token comments of support from a lot of people and a few big games like BF4 that AMD has helped develop will even support it but I have a hard time seeing it become widely adopted.
 
Yeah. This is kind of my problem with the whole thing.

I'm sure we will get token comments of support from a lot of people and a few big games like BF4 that AMD has helped develop will even support it but I have a hard time seeing it become widely adopted.

Yeah. Also if it's an extension, Nvidia will have better chance of implementing their own solution without adding additional burden on game developers. It would be good for the whole industry but I guess it's bad for AMD...
 

atbigelow

Member
Mantle is interesting, but there are huge red flags:

-Interoperability is one of PC's great strengths. Don't just throw it away.

-Do we actually have any benchmarks on Mantle? I know we have the claims about Frostbite 3.

-Smells like 2013's version of Glide. That was a bad time. Let's not go back.
 

Hip Hop

Member
So sometime by the end of 2014?

Because it seems Nvidia currently has a deal with Activision to push Call of Duty and Physx.
 

fallagin

Member
So then, mantle sounds pretty good. Good job amd. I may need to go amd for my next card upgrade unless nvidia is able to come up with similar api somehow.
 

red731

Member
I guess it is hard to learn something new when you are still using idTech3.
(someone correct me if I am wrong)
 
His comments don't make any sense. A GL extension wouldn't accomplish what Mantle is doing, and "wrapping DX or GL around it" would defeat the purpose completely. He must have made this tweet before he knew anything about what Mantle actually is.
 

Contra11

Banned
So .. What Nvidia will do about this Mantle api ? i doubt that they will accept this api that performs better with AMD graphic cards
 

Again...this is NOT a good thing.

Do you really want devs like obsidian and Kerberos using MANTLE (and I'm not sure why OpenGL is considered better..on benchmarks, DirectX generally beats OpenGL

see http://www.g-truc.net/post-0547.html

and this is a guy who pushes OpenGL pretty hard AND as he mentions, this was a test where the testing method (unigine) was OpenGL to begin with....

That sure as hell doesn't look like DrectX is a poor API. (at least compared to OpenGL).)

Unless everyone starts using the big licensed engines Cryengine, Frostbite Idtech, I am absolutely terrified about this future given the average programmer/developer.

Only triple AAA games are going to be able to work with Mantle (even DICE admits that Mantle isn't for everyone and shouldn't be used...that's the whole reason why it links so smoothly with DirectX

One thing in Mantle's favour is that it is NOT Glide. AMD has explicitly said a game coded only for DirectX will work as normal with an AMD directx supproted card and the reverse is also true...

Now..how they accomplished that is WAY, WAY above my understanding...
 

onesvenus

Member
Having it as opengl extension would be much better. It's got more chance of surviving that way.

Yeah. This is kind of my problem with the whole thing.

Doesn't being an OpenGL extension be a little counter productive? Mantle is used to bypass the thickness of OpenGL/DX, how are you suggesting to work as an extension?

Sure you could have calls that interact with the underlying metal more directly through an extension (and there are extensions that do just that) but you'd have to rely on the standard OpenGL way of doing things. I don't think it would be possible to achieve what they claim Mantle can achieve (9x the draw calls for example) without using a lot of extensions to redefine the standard OpenGL calls... and then you get to the point of why doing so instead of defining something completely apart?
 

Codeblew

Member
Again...this is NOT a good thing.

Do you really want devs like obsidian and Kerberos using MANTLE (and I'm not sure why OpenGL is considered better..on benchmarks, DirectX generally beats OpenGL

see http://www.g-truc.net/post-0547.html

and this is a guy who pushes OpenGL pretty hard AND as he mentions, this was a test where the testing method (unigine) was OpenGL to begin with....

That sure as hell doesn't look like DrectX is a poor API. (at least compared to OpenGL).)

Unless everyone starts using the big licensed engines Cryengine, Frostbite Idtech, I am absolutely terrified about this future given the average programmer/developer.

Only triple AAA games are going to be able to work with Mantle (even DICE admits that Mantle isn't for everyone and shouldn't be used...that's the whole reason why it links so smoothly with DirectX

One thing in Mantle's favour is that it is NOT Glide. AMD has explicitly said a game coded only for DirectX will work as normal with an AMD directx supproted card and the reverse is also true...

Now..how they accomplished that is WAY, WAY above my understanding...

Everything is more preferable than DirectX because it is not limited to Microsoft operating systems.
 
-Smells like 2013's version of Glide. That was a bad time. Let's not go back.
Just wanted to say that Glide was an amazing thing for my 12yr old ass w/ shitty integrated graphics that didn't understand anything about video cards and just wanted to play Unreal Tournament. :p it was literally the only way I could run the game and I loved it
 

Truespeed

Member
Again...this is NOT a good thing.

Do you really want devs like obsidian and Kerberos using MANTLE (and I'm not sure why OpenGL is considered better..on benchmarks, DirectX generally beats OpenGL

see http://www.g-truc.net/post-0547.html

and this is a guy who pushes OpenGL pretty hard AND as he mentions, this was a test where the testing method (unigine) was OpenGL to begin with....

That sure as hell doesn't look like DrectX is a poor API. (at least compared to OpenGL).)

Unless everyone starts using the big licensed engines Cryengine, Frostbite Idtech, I am absolutely terrified about this future given the average programmer/developer.

Only triple AAA games are going to be able to work with Mantle (even DICE admits that Mantle isn't for everyone and shouldn't be used...that's the whole reason why it links so smoothly with DirectX

One thing in Mantle's favour is that it is NOT Glide. AMD has explicitly said a game coded only for DirectX will work as normal with an AMD directx supproted card and the reverse is also true...

Now..how they accomplished that is WAY, WAY above my understanding...

And I disagree. The adoption of an open, more efficient cross platform solution is better than what we have right now. It's about time the developers and OEM's were given greater say in the evolution of the the API. AMD has been complaining about the performance issues with DirectX and now the full potential of their hardware can finally be realized. Besides, Microsoft should be glad that this burden has been taken from them. This gives them more time and resources to devote to their OS, devices, etc.
 
Doesn't being an OpenGL extension be a little counter productive? Mantle is used to bypass the thickness of OpenGL/DX, how are you suggesting to work as an extension?

Sure you could have calls that interact with the underlying metal more directly through an extension (and there are extensions that do just that) but you'd have to rely on the standard OpenGL way of doing things. I don't think it would be possible to achieve what they claim Mantle can achieve (9x the draw calls for example) without using a lot of extensions to redefine the standard OpenGL calls... and then you get to the point of why doing so instead of defining something completely apart?

If an api is completely apart from the standards such as opengl, not as many games will adopt it.

Also I highly doubt mantle will be bringing improvement to every opengl calls. It will most likely perform a lot better for a subset of tasks. So if it's implemented as an extension, it'd be easier for coders to maintain different path for different systems that support mantle and those that don't. If both AMD and Nvidia decided to go for their own api's, none of those two will last very long.

Besides, opengl is already a mess with extensions anyways. No reason to hold back now right?
 
And I disagree. The adoption of an open, more efficient cross platform solution is better than what we have right now. .

Except that OpenGL is NOT BETTER.

It loses to DirectX since DirectX9 came out when matched head to head. The documentation and support for OpenGL is also way more haphazard than DirectX.

The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.
 
Except that OpenGL is NOT BETTER.

It loses to DirectX since DirectX9 came out when matched head to head. The documentation and support for OpenGL is also way more haphazard than DirectX.

The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.

Only on Windows. Linux OpenGL performance in valve's testing is better than DirectX.

Also, OpenGL is not open source. It's an open standard. The drivers are actually where the work for OpenGL lies. Therefore, any lack of performance in OpenGL is because the card manufacturer makes poor drivers for OpenGL.
 

Truespeed

Member
Except that OpenGL is NOT BETTER.

It loses to DirectX since DirectX9 came out when matched head to head. The documentation and support for OpenGL is also way more haphazard than DirectX.

The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.

True, open doesn't necessarily always mean better. But, it does allow companies to get involved and start fixing issues that otherwise wouldn't be addressed or would just be put into a queue and looked at when some big company deemed it was in their best interest to modify their proprietary source code. Case in point, Valve were able to get their games running faster on Linux than on Windows. So more involvement by companies in the guidance and development of this new API can only be a good thing and, although, it's true that open isn't always better than closed, I'm willing to bet that the combined talent and knowledge of all of these companies is greater than Microsoft.
 
Except that OpenGL is NOT BETTER.

It loses to DirectX since DirectX9 came out when matched head to head. The documentation and support for OpenGL is also way more haphazard than DirectX.

The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.

Its not that cut and dry at all. For instance, the OpenGL ES variant used on PS3 performs better than the DX counter part.
 

mrklaw

MrArseFace
This could be good news. Apart for you getting more from your GPU than before (and therefore more value for money), this may push drivers to be more reflective of how console drivers work. Maybe it isn't just about power but also about directX not being responsive to developer needs, being too abstracted for efficient game use.

Now, MS must know this - they do the driver work for Xbox 360 and Xbox one. So IMO the best response would be to rework directX 12 or 13 to include mantle-like features. Allow developers more open access to the GPU while still providing the familiarity hat direct X offers.

I don't know why people are worried by mantle. It provides better performance, and shakes things up and encourages nvidia and MS to respond. You'd honestly prefer things were just left how they were? You spooge over a new GPU offering 20-25% more performance costing hundreds of dollars, yet you turn your nose up at someone giving you that for free (hypothetically)
 

Slavik81

Member
The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.
That's a common misconception.
Is OpenGL Open Source?
No, OpenGL doesn't have any source code. GL is a specification which can be found on this website. It describes the interface the programmer uses and expected behavior. OpenGL is an open specification. Anyone can download the spec for free. This is as opposed to ISO standards and specifications, which cost money to access.
There is an implementation of GL that is Open Source and it is called Mesa3D It announces itself as implementing OpenGL 3.0 and GLSL 1.30.
http://www.opengl.org/wiki/FAQ#Is_OpenGL_Open_Source.3F
 

Perkel

Banned
Mantle is interesting, but there are huge red flags:

-Interoperability is one of PC's great strengths. Don't just throw it away.

-Do we actually have any benchmarks on Mantle? I know we have the claims about Frostbite 3.

-Smells like 2013's version of Glide. That was a bad time. Let's not go back.

Bullshit. Glide was awesome.
 

wsippel

Banned
Except that OpenGL is NOT BETTER.

It loses to DirectX since DirectX9 came out when matched head to head. The documentation and support for OpenGL is also way more haphazard than DirectX.

The biggest and ONLY reason openGL is used is because it's open source.

Open Source does not automatically mean "BETTER" than proprietary.
OpenGL is the standard API for professional applications. Professionals value stability and precision over raw performance. Therefore, the drivers are optimized to deliver just that, not for high performance gaming. Nobody in his right mind would use D3D to drive something like Solidworks or Nuke.
 

Slavik81

Member
I don't know why people are worried by mantle. It provides better performance, and shakes things up and encourages nvidia and MS to respond. You'd honestly prefer things were just left how they were? You spooge over a new GPU offering 20-25% more performance costing hundreds of dollars, yet you turn your nose up at someone giving you that for free (hypothetically)
It's not free. The price is paid by developers who must write, test and maintain additional code paths. Though, they have to support a few paths already anyway.

I'm just hoping the API is cheap and simple to support, because it's definitely not going to replace other APIs any time soon.
 

OmegaDL50

Member
Anything that allows the developers to get access to lower level hardware without more overhead and API's getting in the way of maximizing performance is always a good thing.

Console development is done in such a fashion not having to deal with the issue of OS overhead, which means they can dedicate more of the systems hardware to running the game in an efficient manner. If they can do the same on PC, then those that game on PC only serve to benefit from this.

Hopefully this becomes something that reduces hardware requirements overall. I'll be looking forward to see how this develops.
 

onesvenus

Member
If an api is completely apart from the standards such as opengl, not as many games will adopt it.
Well, that's until it becomes a standard in itself.

Also I highly doubt mantle will be bringing improvement to every opengl calls. It will most likely perform a lot better for a subset of tasks. So if it's implemented as an extension, it'd be easier for coders to maintain different path for different systems that support mantle and those that don't. If both AMD and Nvidia decided to go for their own api's, none of those two will last very long.

Besides, opengl is already a mess with extensions anyways. No reason to hold back now right?
That's a poor argument. Having utilized OpenGL and DirectX for a long time I can see how a new standard can be beneficial. Of course it would come with its own set of drawbacks but mantaining an API for so long what it usually does is add a lot of problems in its own. If you add that the underlying architecture has changed a lot from when those APIs were defined (and yes, I know that they have been updated but their workings are more or less the same) it can be a slow layer compared to what could be done today having the current architectures as a base.
To compete OpenGL and DirectX would have to be redefined and lose the backwards compatibility with older architectures and I'm sure no one has the balls to do it. The only solution is AMD/Nvidia or another player to create its new standard that is much more directly tied with the current hardware.
 
Top Bottom