AMD has given Khronos unfettered access to Mantle

[monstercameron]

 

 

Huddy told us AMD has done a "great deal of work" with the Khronos Group, the stewards of the OpenGL spec, on OpenGL Next. AMD has given the organization unfettered access to Mantle and told them, in so many words, "This is how we do it. If you want to take the same approach, go ahead." Khronos is free to take as many pages as it wants out of the Mantle playbook, and AMD will impose no restrictions, nor will it charge any licensing fees.

source: http://techreport.com/news/26922/amd-hopes-to-put-a-little-mantle-in-opengl-next

 

I mean who didnt see this coming?

[BallGum]

source: http://techreport.com/news/26922/amd-hopes-to-put-a-little-mantle-in-opengl-next

 

I mean who didnt see this coming?

 

Many of the people who shouting that Mantle isn't open. Technically, they were right, it isn't really open yet, but we will hopefully get there, and this news is very promising.

[Askew]

So can someone explain to me what this means for OpenGL going forward?

 

Is this AMDs way of opening up the Mantle API?

[Askew]

AMD have always been a considerably open source friendly company. Intel have been open friendly as well. nVidia on the other hand is another story. 

 

gameworks.dll anyone? lol

[helping]

So can someone explain to me what this means for OpenGL going forward?

 

Is this AMDs way of opening up the Mantle API?

It mainly means that Nvidia will be able to dick around with it for giggles and research, they'll probably use it to improve their technology.

 

It's in AMD's best interest regardless, best case scenario it helps Mantle catch on. 

[BallGum]

So this could trickle down to Android, which is cool. I can't wait for more mobile devices to start being used as consoles as well.

[Askew]

This mainly means that all the companies within the Khronos group will potentially have access to this code, even nVidia, to help build a better OpenGL that can be used on all sets of hardware from our beautiful desktop GPUs, to tablet and phone chips and even fully embedded systems used in industry.

It mainly means that Nvidia will be able to dick around with it for giggles and research, they'll probably use it to improve their technology.

 

It's in AMD's best interest regardless, best case scenario it helps Mantle catch on. 

 

So D3D could be closer to death than it was yesterday, and SteamOS could be closer to hosting a unified API on Linux than it was yesterday?

 

if so, tres gustavo, I've got the impression of late that reliance on D3D is a negative for PC gaming.

[patrickjp93]

I know AMD snubbed Intel's request to play around with it to give themselves as much of an edge as possible in the oncoming iGPU war (begins in earnest with Skylake), but everyone knows they will either have to keep it closed (pissing off their own base) or give Intel access (hurting their competitiveness in markets they'd like to break into, such as scientific computing and laptops where they barely have a presence).

[monstercameron]

I know AMD snubbed Intel's request to play around with it to give themselves as much of an edge as possible in the oncoming iGPU war (begins in earnest with Skylake), but everyone knows they will either have to keep it closed (pissing off their own base) or give Intel access (hurting their competitiveness in markets they'd like to break into, such as scientific computing and laptops where they barely have a presence).

a bit misguided as AMD just doesn't have the market share to do this!

[patrickjp93]

a bit misguided as AMD just doesn't have the market share to do this!

But it wants the market share, and it's caught between a rock, a hard place, and TSMC/GloFo production processes.

[bogus]

Great! Lets just hope NVIDIA doesn't pressure them to not accept in exchange for a ridiculous ammount of money.

[monstercameron]

But it wants the market share, and it's caught between a rock, a hard place, and TSMC/GloFo production processes.

tsmc and co. arent an issue...AMD could be just as competitive even on the old 32nm process. The problem is their marketing and partnerships. The performance of the chip only matters so much.

[BallGum]

I know AMD snubbed Intel's request to play around with it to give themselves as much of an edge as possible in the oncoming iGPU war (begins in earnest with Skylake), but everyone knows they will either have to keep it closed (pissing off their own base) or give Intel access (hurting their competitiveness in markets they'd like to break into, such as scientific computing and laptops where they barely have a presence).

 

I literally see you on every forum post regarding either nVidia or AMD, and you're always advocating / defending nVidia or downplaying AMD. Forgive me if I think you have some sort of agenda.

[patrickjp93]

tsmc and co. are an issue...AMD could be just as competitive even on the old 32nm process. The problem is their marketing and partnerships. The performance of the chip only matters so much.

Performance and performance/watt are everything in the server/corporate market. Without process shrinks they can't add transistors, which means logic can't improve beyond a certain point, which means Intel can sit on its laurels in terms of implementation and can spend money on finding solutions to implement later if AMD becomes a threat with ARM or its own x86 chips in any of their big markets. That's why Intel went on the fast track to going heterogeneous and catches up to Kaveri with Skylake, and it will doubtlessly match Carrizo with the successor to Cannonlake, and that will be on a 10nm process instead of AMD's 28/20nm process.

[bogus]

Oh I tryed to communicate with that fellow - according to him AMD brings no inovation, just lags everything behind and NVIDIA pretty much is some sort of god.

[patrickjp93]

I literally see you on every forum post regarding either nVidia or AMD, and you're always advocating / defending nVidia or downplaying AMD. Forgive me if I think you have some sort of agenda.

I'm a realist from a business perspective. Nvidia wants to defend its proprietary work and make you pay for it. That's their prerogative. AMD puts everything at a lower price point, making less profit per chip, but it gets wider consumer support which helps make up some of the difference.

 

As per AMD v. Intel, I want AMD to come out and punch Intel in the face, but it's not going to happen because AMD sold GloFo and has very little control over process shrinks (which it needs if it's going to compete with Intel, because there's only so much you can squeeze from architecture alone within a certain transistor budget.

 

I don't have an agenda other than the truth and seeking out cool innovations.

[monstercameron]

Performance and performance/watt are everything in the server/corporate market. Without process shrinks they can't add transistors, which means logic can't improve beyond a certain point, which means Intel can sit on its laurels in terms of implementation and can spend money on finding solutions to implement later if AMD becomes a threat with ARM or its own x86 chips in any of their big markets. That's why Intel went on the fast track to going heterogeneous and catches up to Kaveri with Skylake, and it will doubtlessly match Carrizo with the successor to Cannonlake, and that will be on a 10nm process instead of AMD's 28/20nm process.

even if amd had a 5nm quantum processor, unless they could build meaningful partnerships with odms and oems it doesnt matter.  Superior tech fails all the time.

[Sky Daddy]

I know AMD snubbed Intel's request to play around with it to give themselves as much of an edge as possible in the oncoming iGPU war (begins in earnest with Skylake), but everyone knows they will either have to keep it closed (pissing off their own base) or give Intel access (hurting their competitiveness in markets they'd like to break into, such as scientific computing and laptops where they barely have a presence).

Not exactly. 

 

AMD didn't give access to Intel because it's not out of beta yet. Intel is not developing their own api either so they can't possibly gain as much as khronos has with it. 

Mantle will basically be a failure as a closed standard. Intel has the more gpus than amd and nvidia combined. That being said, games developed on mantle could reach a wider audience = more money if intel supports it. 

[deathjester]

Great! Lets just hope NVIDIA doesn't pressure them to not accept in exchange for a ridiculous ammount of money.

 

Nvidia is a Steam OS partner. You are looking at the wrong company. Nvidia just supports it all and will go with whatever remains. Add to that Nvidia is a Kronos partner. AKA Nvidia/ATI/Intel OWN OpenGL. 

 

DirectX 12 = Mantle = Mantle/OpenGL. Shader library is different but neither is ever behind. The porting between Mantle and DX 12 is supposedly easy as hell. 

 

http://www.eteknix.com/amd-demonstrates-mantle-can-be-easily-ported-to-directx-12/

 

There is only one company who wants to stop OpenGL adoption and it is not Nvidia. Kronos getting access to Mantle means Nvidia now has access to Mantle.

 

Nvidia could possibly make more money with gaming computers costing 100 less dollars (no Windows). Gaming computers will be cheaper. MS is who stands to lose money here, not Nvidia.

 

http://en.wikipedia.org/wiki/Khronos_Group

 

Look at the partners. Intel is also a partner. That means Intel now has AMD Mantle as well. Looks like AMD was not lying. Mantle IS open source.

[patrickjp93]

Oh I tryed to communicate with that fellow - according to him AMD brings no inovation, just lags everything behind and NVIDIA pretty much is some sort of god.

PFFFT! Obfuscation and slander.

AMD hasn't done anything truly impressive since it went dual-core or created a flawed 64-bit instruction set which beat Intel to market and gained popularity purely by ease of use (performance drawbacks aside). Itanium was then rushed to market and was a hot mess. Now we're stuck with AMD-64 and its limitations until either company completely rethinks it and makes a better model (and the cross-compiler to make the transition easy).

 

As per Mantle, I'm a bit miffed AMD just swiped up the original programmers and took the project out of the open source world, and I don't buy that they'll make it open until it is, because from a business analyst's perspective, it's stupid of them to do it with $2 billion in debt hanging over their heads while not being able to innovate their chip design as much as they need to.

 

Though, Nvidia will be the first of the great triad to fall unless Tegra can take off in a way that can seriously compete with Broadwell. They're not good at making x86 or ARM chips, which leaves them vulnerable unless they have a truly otherworldly GPU design waiting for when iGPUs become so powerful it needs a way to justify its expense to gamers and supercomputer builders.

[patrickjp93]

even if amd had a 5nm quantum processor, unless they could build meaningful partnerships with odms and oems it doesnt matter.  Superior tech fails all the time.

Superior tech fails all the time usually because it's too expensive. AMD sells cheap and the world knows it. All they need is a competitive product.

[patrickjp93]

Not exactly. 

 

AMD didn't give access to Intel because it's not out of beta yet. Intel is not developing their own api either so they can't possibly gain as much as khronos has with it. 

Mantle will basically be a failure as a closed standard. Intel has the more gpus than amd and nvidia combined. That being said, games developed on mantle could reach a wider audience = more money if intel supports it. 

No no no, that would threaten AMD's APUs and GPUs alike. They already have enough partnerships with game studios to not need to release it to open waters and make their GPUs competitive against Nvidia's in the gaming world (and in the scientific world loathe though I am to admit it, CUDA being a far easier and superior programming platform if you know C/C++). AMD wants the gold, and it might mean sacrificing some base happiness while attracting more buyers from elsewhere who want a cheaper alternative to Intel (or cheaper Intel).

[patrickjp93]

Nvidia is a Steam OS partner. You are looking at the wrong company. Nvidia just supports it all and will go with whatever remains. Add to that Nvidia is a Kronos partner. AKA Nvidia/ATI/Intel OWN OpenGL. 

 

DirectX 12 = Mantle = Mantle/OpenGL. Shader library is different but neither is ever behind. The porting between Mantle and DX 12 is supposedly easy as hell. 

 

http://www.eteknix.com/amd-demonstrates-mantle-can-be-easily-ported-to-directx-12/

 

There is only one company who wants to stop OpenGL adoption and it is not Nvidia. Kronos getting access to Mantle means Nvidia now has access to Mantle.

 

Nvidia could possibly make more money with gaming computers costing 100 less dollars (no Windows). Gaming computers will be cheaper. MS is who stands to lose money here, not Nvidia.

 

http://en.wikipedia.org/wiki/Khronos_Group

 

Look at the partners.

That's what I like, good business analysis. Just wait until game programmers all start going for Linux distributions (is steam based on a Linux/Unix flavor?) where Microsoft could lose out to open office (if ever that ugly ass UI is fixed).

[monstercameron]

snip

libre office!

[patrickjp93]

libre office!

One flavor of Open Office. Still fugly.