[PcPer] Some mantle news from GDC

[tviceman]

The shortest lived, over-hyped API ever.

 

[Silverforce11]

How do you retire a product that never made it out of beta? Cancelled is the more accurate term.

Depends whether you think Mantle is the foundation of DX12 or not.

I do. DICE does. Many studios who signed on with the Mantle SDK did it to prepare for DX12 also think so.

To me, Mantle lives on in DX12, particularly if DX12 indeed brings asynchronous DMA & compute, that will lead to major rendering improvements under GPU bottleneck situations, on top of "Mantle 1.0" which is the draw-call overhead reduction.

 

[SimianR]

It definitely seemed to drive some features into DX12. It's just too bad that AMD will have one less marketing bullet point on their boxes. How can you compete when the competition has "OVER THE TOP SMOKE AND RAGDOLL EFFECTS THROUGH physX!".

 

[DiogoDX]

Don't worry guys. Mantle 1.0 just get a new name and Khronos will release the specs.

The codename "Vulkan" is indeed highly suggestive.:hmm:

 

[frozentundra123456]

We actually don't know that. For all we know Mantle 1.0 could have been finished months ago, but with it being under NDA there is just no way of knowing that.

I guess we will know more later this week.

Well, either it never made it out of "beta", or AMD was lying, because they said the reason it was not released publicly is that it was still in beta. Although I have said in earlier forums that it made no sense to me why they could not release the beta version if they truly wanted it to be open.

 

[Noctifer616]

Well, either it never made it out of "beta", or AMD was lying, because they said the reason it was not released publicly is that it was still in beta. Although I have said in earlier forums that it made no sense to me why they could not release the beta version if they truly wanted it to be open.

A product being open or not has nothing to do with it's development state. I am merely saying that we can't know the exact state of the Mantle SDK as long as it's limited to devs only.

We might know more from the Mantle paper they will release this month. We will see.

 

[blastingcap]

It definitely seemed to drive some features into DX12. It's just too bad that AMD will have one less marketing bullet point on their boxes. How can you compete when the competition has "OVER THE TOP SMOKE AND RAGDOLL EFFECTS THROUGH physX!".

It's ok, AMD can afford to have wasted so much time/effort/money in Mantle development because AMD is a rich company rolling in dough...

Oh wait. Whoops.

Is it too late for ATI to get a divorce? Getting married to AMD has been awful for ATI.

 

[monstercameron]

Can't say I'm not disappointed, amd should follow through on their promises...then again when your expectations of amd are so low, this doesn't even scratch the surface.

I hold out a little hope that Vulcan will be related to this decision.

 

[mindbomb]

So Mantle is officially dead. DX12 won on TKO.

they basically used mantle as leverage to force dx12 to adopt features they liked.

 

[Spjut]

I find it hard to believe Mantle isn't very similar to either DirectX 12 or Vulkan.

I just don't believe AMD would spend resources on its own API and expect it to become standard, and then just drop it like this. AMD is also obviously involved in both DirectX and OpenGL.

 

[SimianR]

It's ok, AMD can afford to have wasted so much time/effort/money in Mantle development because AMD is a rich company rolling in dough...

Oh wait. Whoops.

Is it too late for ATI to get a divorce? Getting married to AMD has been awful for ATI.

Yup, if Mantle is truly dead then it benefited everyone for pushing the same features into DX12 but didn't do much for cash strapped AMD.

 

[destrekor]

Don't worry guys. Mantle 1.0 just get a new name and Khronos will release the specs.

This very much sounds like it will be the case.

Mantle’s definition of “open” must widen. It already has, in fact. This vital effort has replaced our intention to release a public Mantle SDK, and you will learn the facts on Thursday, March 5 at GDC 2015.

This simply speaks to the idea that they are now moving from an "open" project that has a publicly-available APK, to an open project that will be at least partially open to public development efforts. This jives strongly with the intent of OpenGL and what may be glNext, with Vulkan appearing to be the new name for that project.

Then there is this, which indicates AMD has been open to the new OpenGL drawing from the Mantle body of work, in whatever way Khronos Group saw fit.

The Thursday schedule at GDC as it relates to these topics is thus:

AMD's only talks that day, one is a TressFX 3.0 talk for Deus Ex Universe Projects, and the other is showcasing DirectX 12 efficiency and performance, with a Microsoft guest speaker

Valve is presenting glNext. One of the speakers? Johan Andersson, of EA and the Frostbite Engine team.

That last bit is pretty telling, IMHO. Frostbite has never been anything but DirectX and Mantle. They've never developed for OpenGL, and with already using TWO APIs, why would they tack on a third? Well, one possibility could be they are just replacing Mantle with glNext, end of story, but it could be that they are only even considering that because they've already laid the foundation and glNext is going to basically be Mantle 1.0 or at least draw very heavily from Mantle.

I feel confident in arguing that AMD was fully aware that Mantle would never be a long-term solution. They took what they knew of current DirectX 12 tech goals, and took what they learned and helped create for the Xbox One with Microsoft, and turned that into a PC API for their architecture, and took the time to include overhead reduction strategies and more "bare metal", which sounds like was not a big push of the original DirectX 12 development.

Regardless, they made an entrance as the THIRD API of importance on the PC: DirectX, OpenGL, and now Mantle. The world is hardly big enough for two. Three? I think AMD fully understood that Mantle would never "dethrone" OpenGL, frankly they could never even dream of competing with OpenGL in terms of developer mindshare or market presence.

That is why they offered up Mantle to Khronos Group last summer. Mantle was basically a public demonstration of a proof of concept, and AMD might just want Mantle to be pushed further to include more architectures and more platforms.

edit:

Heck, I think that Khronos Group has chosen "Vulkan" as the likely brand name for glNext is a pretty telling fact. Vulkan so heavily references the theme of Mantle that it seems perfectly intentional, and may strongly suggest how close the code and objective of Vulkan is to that of Mantle.

edit2:

Someone on Reddit said this, which is perfect as an image:
"If the Earth (Mantle) opens up you get a Volcano. Volkan -> OpenMantle."
It was previously mentioned that Volkan means Volcano in numerous languages.

 

[DrMrLordX]

Valve is presenting glNext. One of the speakers? Johan Andersson, of EA and the Frostbite Engine team.

Ah hah. The first thing I was wondering when I read this thread was: what's this going to do to Valve? Valve is making a pretty big push to get gaming off the Windows desktop platform with SteamOS. Mantle is something they could have used in some of their own Linux games, and other Steam developers could have done the same. If it lives on in the form of Vulkan/glNext, then they can salvage what is left of Mantle and use it for their own purposes.

 

[ocre]

after thinking about thinking about this for hours and now i understand that this was always the plan from amd from day one.

they actually single handedly created the one and only API of the future, for the benefit of all of mankind but knew that leaving it named mantle would only prevent 80% of PC gamers from ever having the chance to be changed for life because evil nvidia would block mantle from ever being used on their cards.

So AMD sacrifices mantle.........but that's just what they want nvidia to think. See, all the signs are there,

Vulcan and mantle, they both have 6 letters. Think about it? The odds of that are astronomical and it is really just a sign to reach those who can see.

Then there is another clue, which describes the true reason behind all of this.

MANtle + vulCAN

MAN-CAN.

Its the secret message to describe the true API of the future, the ultimate token to unite all gamers under one perfect API of all time.

Man can.
Nothing can stop the will of man when he really sets his mind to it. Not ocean,, not sea, and not even space. Man can prevail and unite. The world comes together under a single most perfect API, the API of the future.

AMD created Vulcan so nvidia users would finally experience the sophistications of a modern api, it is mantle for everyone. But they knew as long as it was attached to AMD, people would refuse it, run for the hills to get as far as they could away from it. So for the betterment of man kind, now man can experience the API of the future. All of man can.

Btw, this is a light hearted joke. So please no one get offended.

 

[gdansk]

I was hoping AMD's Mantle push would be good for PC gaming. D3D12 and glNext seem to be taking up the mantle such that AMD no longer needs to.

I wonder how long until glNext is on phones/tablets.

 

[destrekor]

after thinking about thinking about this for hours and now i understand that this was always the plan from amd from day one.

they actually single handedly created the one and only API of the future, for the benefit of all of mankind but knew that leaving it named mantle would only prevent 80% of PC gamers from ever having the chance to be changed for life because evil nvidia would block mantle from ever being used on their cards.

So AMD sacrifices mantle.........but that's just what they want nvidia to think. See, all the signs are there,

Vulcan and mantle, they both have 6 letters. Think about it? The odds of that are astronomical and it is really just a sign to reach those who can see.

Then there is another clue, which describes the true reason behind all of this.

MANtle + vulCAN

MAN-CAN.

Its the secret message to describe the true API of the future, the ultimate token to unite all gamers under one perfect API of all time.

Man can.
Nothing can stop the will of man when he really sets his mind to it. Not ocean,, not sea, and not even space. Man can prevail and unite. The world comes together under a single most perfect API, the API of the future.

AMD created Vulcan so nvidia users would finally experience the sophistications of a modern api, it is mantle for everyone. But they knew as long as it was attached to AMD, people would refuse it, run for the hills to get as far as they could away from it. So for the betterment of man kind, now man can experience the API of the future. All of man can.

Btw, this is a light hearted joke. So please no one get offended.

I know that was in jest, but FYI, it is properly trademarked "Vulkan" so that throws off that little secret message a tad.

Additionally, nVidia could never implement Mantle if they even wanted to, the API is directly written for the benefit of only the GCN micro-architecture. As it is directly written, it is useless to attempt to use it elsewhere.

That said, the idea behind Mantle is solid, and if the next version of OpenGL is essentially a platform-agnostic version of Mantle, well, that's basically a win for everyone.

To be fair, I hope they really do well, because I was frankly never too pleased with OpenGL in the past. Engines that utilized it always seemed to have this "character" to them that I was never fond of.

Then again, perhaps I simply hate the appearance of iD Tech engines. I haven't checked out anything like CS or Half Life in an OpenGL environment. I'll have to do that someday, load up Steam under OS X once I have it working again on my desktop, see how that engine under OpenGL compares to the DX version.

As for Nvidia, they'll have plenty of ways to utilize the same kind of power as Mantle through DX12. I don't expect Vulkan to have much on DX12 as a universal API, but I wonder if there will be any vendor-specific coding to help games coded for Vulkan perform the same on GCN as one would under Mantle. Obviously Nvidia wouldn't have the same "comparison" possible, but vendor-specific instructions may yet be possible. I have no idea if any of this will and or can even happen at the API level for the universal API types. Actually, that may rest on the shoulders of game developers if the API is sufficiently closer to the metal.

I guess, in the end, regardless of how it all shapes up, we all win. Both of the old mainstays, DirectX and OpenGL, will be giving the entire gaming community what AMD had managed to give their users a little bit early.

I suspect Microsoft had their hands in the mix when it came to developing the bare metal approach, equally alongside AMD, but even as a fan of Nvidia products (who was "forced" into buying AMD thanks to the 970 fiasco and an absurd level of impatience ), it does seem to me that AMD has directly and significantly contributed to both DX12 and Vulkan, and is to be thanked for helping to push this development style to the forefront in the coming years.

I wonder how well Vulkan will translate to current GPUs. Much like DX12 and previous iterations of OpenGL, not all GPUs will be compatible. Will there be software emulation for different features that an architecture cannot handle natively? Will we have to wait until the next GPUs are available that are specifically designed to natively support Vulkan?

It seems 2015 is shaping up to be a pretty significant year for PC gaming, at least it will be seen that way in history. Whether we have anything by the end of the year to take advantage of all this remains to be seen, but having it available to developers is a major step forward from the era of abstraction-level APIs. To finally be reaching non-vendor-specific "bare metal" APIs is a major step forward, and now there is the groundwork to improve upon the formula and perfect the APIs.

 

[ShintaiDK]

Mantle is now GLNext? Will the BS never stop? :thumbsdown:

 

[BFG10K]

I said right from the start that Mantle would go nowhere. The only people who didn't see this coming were those that didn't want to see.

 

[dacostafilipe]

Will the BS never stop? :thumbsdown:

That's kinda funny coming from you :sneaky:

It's almost certain that Mantle will be used in Vulkan.

Phoronix confirmed that Vulkan will not be compatible with OpenGL, this almost refutes the AZDO idea (AZDO requires OpenGL).

Some AMD devs even confirmed that Khronos has Mantle code for a long time.

Mantle was the catalyst needed to push the industry and it will still live as, multi-vendor, open standard API in the future.

 

[ShintaiDK]

That's kinda funny coming from you :sneaky:

It's almost certain that Mantle will be used in Vulkan.

Phoronix confirmed that Vulkan will not be compatible with OpenGL, this almost refutes the AZDO idea (AZDO requires OpenGL).

Some AMD devs even confirmed that Khronos has Mantle code for a long time.

Mantle was the catalyst needed to push the industry and it will still live as, multi-vendor, open standard API in the future.

There is no DX12, only Mantle..Oh no.

Mantle is the console API..Oh no.

Mantle is DX12..Oh no.

Mantle is Open..Oh no.

Mantle isnt tied to GCN..Oh no.

Mantle wont be replaced by DX12..Oh no.

Mantle is now GLNext!

The blind fans simply keeps shifting the goalpost.

 

[dacostafilipe]

...
The blind fans simply keeps shifting the goalpost.
...

I don't think this even requires an answer, does it? :hmm:

Anyway, Vulkan was just announced : https://www.khronos.org/assets/uploads/developers/library/overview/2015_vulkan_v1_Overview.pdf

 

[nvgpu]

http://www.anandtech.com/show/9038/next-generation-opengl-becomes-vulkan-additional-details-released

and though we need to be clear that Vulkan is not Mantle

What has changes is that Khronos has gone through a period of refinement, throwing out portions of Mantle that didn’t work well in Vulkan – mainly anything that would prevent it from being cross-vendor – and replacing it with the other necessary/better functionality.

Vulkan is not Mantle.

Mantle was dead on arrival from the start, just accept it.

 

[dacostafilipe]

Specially for ShintaiDK

Here's a post about Vulkan from Imagination : http://blog.imgtec.com/powervr/trying-out-the-new-vulkan-graphics-api-on-powervr-gpus

There's also some Vulkan source code :

vkCmdBindDescriptorSet(cmdBuffer, VK_PIPELINE_BIND_POINT_GRAPHICS,
textureDescriptorSet[0], 0);
vkQueueSubmit(graphicsQueue, 1, &cmdBuffer, 0, 0, fence);
vkMapMemory(staticUniformBufferMemory, 0, (void **)&data);
// ...
vkUnmapMemory(staticUniformBufferMemory);

Now, let's open mantle64.dll and get the list of methods:

http://stackoverflow.com/questions/20966317/does-amd-mantle-api-already-available

There we find:

grCmdBindDescriptorSet
grQueueSubmit
grMapMemory
grUnmapMemory

 

[ShintaiDK]

You cant be serious. Just because they got similar names you think they are the same API and code?

 

[BFG10K]

There is no DX12, only Mantle..Oh no.

Mantle is the console API..Oh no.

Mantle is DX12..Oh no.

Mantle is Open..Oh no.

Mantle isnt tied to GCN..Oh no.

Mantle wont be replaced by DX12..Oh no.

Mantle is now GLNext!

The blind fans simply keeps shifting the goalpost.

Also:

*Soon Mantle will be out of beta and the SDK will be available.

*Mantle will be a multi-vendor API because Intel asked to look at the documents.

I'm reading the older Mantle threads and chuckling at some of the comments there.