AMD GPU14 Tech Event Sept 25 - AMD Hawiian Islands

[Bubbleawsome]

The prices are the same, AMD is rebranding their current lineup and sticking two new GPUs at the top to compete with 780 and Titan.

So is it wort waiting for the 280/290. And I'm wondering what the dual GPU will be. 299? 292? 290x2?

 

[RiDE]

Judging by the feathers being ruffled in this thread you can almost see who's biased for whom. :hmm:

I'm for anything that might help linux become a better platform for pc gaming in the long run.

 

[Saylick]

AMD >>better<< not use GCN AGAIN at 20nm.

If they aren't using GCN for their 20nm GPUs, they won't get any benefit from this stuff for more than literally half a year give or take.

Trying to fight a new architecture like Maxwell with GCN AGAIN is just madness.

I'm sure they will make revisions to GCN to ensure it remains competitive with the competition. Similarly, if this whole Mantle thing pans out, they will update it to be compatible with future revisions of GCN as well.

 

[Silverforce11]

AMD >>better<< not use GCN AGAIN at 20nm.

If they aren't using GCN for their 20nm GPUs, they won't get any benefit from this stuff for more than literally half a year give or take.

Trying to fight a new architecture like Maxwell with GCN AGAIN is just madness.

GCN is being constantly evolved, we've had version 1.1 and soon getting version 2.0, each microarch lasts for many years.

 

[BD2003]

You're way over doing this. It's not like developers are suddenly going to skip developing with DirectX or OpenGL. Seriously.

We've seen this story before, and we know how it ends. Either the proprietary API fails and ends up as a footnote in history (glide). Or it closes off competition, and we get half-assed refreshes of the same technology for a few years (EAX), until the entire industry just gives up and goes another way.

The only people that should be cheering this on are AMD stockholders. This is a huge step back for gamers. At best it's a distraction for a few years, but it could get really ugly if it succeeds. You can call people crazy if you want, but we've seen this happen before, and the results speak for themselves.

Even if they don't stop developing for DX anytime soon, this could give AMD cards such a price/performance advantage, that they just won't be competitive. If that happens, and AMD no longer has NVIDIA lighting a fire under their ass, innovation and performance increases will slow to a crawl.

 

[EightySix Four]

It will work, it will just be slower, thats exactly the point, since NV GPUs running these future games will go through DX and windows.. on GCN, it will go through Mantle. NV will need a massive hardware grunt advantage to compete, since they are stuck with an inefficient API. Efficiency matters. <<-- this is the point of having a low lvl API for console/PC development.

Right, but this is a relative performance increase for AMD cards over nVidia cards, not an absolute decrease for nVidia cards. Your nVidia card that was going to get 45fps in Battlefield 4 is still going to get 45fps in Battlefield 4, it's just that owners of the AMD card who were going to get 45fps may now get 50fps (pulling numbers out of my ass).

I don't really understand where nVidia card owners have a right to be angry about it? There card is just as good as it was, it has only lost performance relative to the competition.

We've seen this story before, and we know how it ends. Either the proprietary API fails and ends up as a footnote in history (glide). Or it closes off competition, and we get half-assed refreshes of the same technology for a few years (EAX), until the entire industry just gives up and goes another way.

The only people that should be cheering this on are AMD stockholders. This is a huge step back for gamers. At best it's a distraction for a few years, but it could get really ugly if it succeeds. You can call people crazy if you want, but we've seen this happen before, and the results speak for themselves.

This is nothing like the previous times precisely because there are already well supported common APIs and multiple competitors in the market place. OpenGL and DirectX won't go anywhere, and Intel and nVidia will continue to chug ahead. It would be moronic to release a game which only support Mantle or even favor Mantle too much. The vast majority of computers out there don't even have a discreet graphics card and use built in Intel graphics...

 

[Keysplayr]

Is that how you feel about PhysX also?

I was just about to ask sushiwarrior the same exact thing!!!! Great minds!! eh?

Quote: "Yes, AMD spent money and man-hours to gain an additional performance benefit for their cards through low level optimization tweaks. Said API is open to Nvidia if they would like to go through the same process as AMD has. But are you suggesting that AMD should use THEIR time and THEIR money to optimize performance on Nvidia cards? LOL! "

 

[flopper]

If I made a game and I had the chance to add low level access support for the game with mantle, and add 2x performance to any gcn card, I jump at the chance.
Now the game bundle and such all make sense.

 

[OCGuy]

Right, but this is a relative performance increase for AMD cards over nVidia cards, not an absolute decrease for nVidia cards. Your nVidia card that was going to get 45fps in Battlefield 4 is still going to get 45fps in Battlefield 4, it's just that owners of the AMD card who were going to get 45fps may now get 50fps (pulling numbers out of my ass).

I don't really understand where nVidia card owners have a right to be angry about it? There card is just as good as it was, it has only lost performance relative to the competition.

Anytime the homogenization of console and PC games is served as a sh*t-sandwich, (especially when graphics are involved), all PC gamers should be concerned.

 

[sontin]

First: I agree with you that it is like CUDA. CUDA is a perfectly excellent tech developed by NV, and is extremely useful.
Second: It's not an artificial lockdown. You need to use the proper terminology. An "artificial" lockdown means it was not needed. CUDA, for example, is not an artificial lockdown. It target's NV's hardware, because it is meant to target a specific architecture, because that's the fastest way to achieve the most performance out of your hardware and arch. So NV needed to do no such artificial lockdowns to make CUDA non-workable in AMD or Intel tech, artificial lockdowns meaning a GPU Vendor ID check, or a GPU Model ID check to see if it is your own company's model, etc. There's no need for that because, by nature of targeting a specific hardware arch, it will automatically not work (fail spectacularly) on a different class of hardware arch.

nVidia only needs to allow AMD to write their own back-end for Cuda. But they will not do it. It's the best example of artificial.

So, no artificial lockdowns. But yes, inherently, it will end up locked to GCN.

Mantle was developed for GCN. It's a specific low level API for a certain architecture. It's not able to run on other hardware. That is a artificial lockdown. Dice can use DX or OpenGL and trying to convince Microsoft for a better support in DX. But they decided to create an own API for GCN. Really that is right the definition of artificial and excactly the same like Cuda.

What won't happen, for sure, is that Mantle supplants DX. That's crazy talk. So it's not the end of the sane world of PC gaming. If that is your fear, then don't worry. The chances of that are nil.

So, why would a developer use Mantle if they still need another DX path? The reason for low level API is that you dont want to optimize for other plattforms. More APIs mean more work, more investment. Mantle makes only sense if you start to ignore other vendors.

And there is no cross plattform development between Mantle and X1 and PS4. Both using their own software stack.

 

[Freaksterz]

I was just about to ask sushiwarrior the same exact thing!!!! Great minds!! eh?

Quote: "Yes, AMD spent money and man-hours to gain an additional performance benefit for their cards through low level optimization tweaks. Said API is open to Nvidia if they would like to go through the same process as AMD has. But are you suggesting that AMD should use THEIR time and THEIR money to optimize performance on Nvidia cards? LOL! "

Except that Physx could be loaded to the CPU just as it was in BL and later got patched.
Here they make new code to make better use of their hardware present in Radeon GPUs, Hardware in PS4 and XB1.

 

[AnandThenMan]

We've seen this story before, and we know how it ends. Either the proprietary API fails and ends up as a footnote in history (glide). Or it closes off competition, and we get half-assed refreshes of the same technology for a few years (EAX), until the entire industry just gives up and goes another way.

The only people that should be cheering this on are AMD stockholders. This is a huge step back for gamers. At best it's a distraction for a few years, but it could get really ugly if it succeeds. You can call people crazy if you want, but we've seen this happen before, and the results speak for themselves.

Even if they don't stop developing for DX anytime soon, this could give AMD cards such a price/performance advantage, that they just won't be competitive. If that happens, and AMD no longer has NVIDIA lighting a fire under their ass, innovation and performance increases will slow to a crawl.

Agree with a lot of this. Understand why AMD is going in this direction, but long term for gamers it's harmful. Problem is, DirectX is just really inefficient (not to mention Windows) so from that perspective I welcome anything that alleviates the wasted processor cycles.

The perfect solution to me would be a close to the metal, open graphics API on Linux. And no OpenGL is not good enough, in fact it's a bit of a mess the last few years.

 

[24601]

If I made a game and I had the chance to add low level access support for the game with mantle, and add 2x performance to any gcn card, I jump at the chance.
Now the game bundle and such all make sense.

If you were making a game and didn't design the entire engine yourself, the choice is not up to you........ at all.

 

[EightySix Four]

nVidia only needs to allow AMD to write their own back-end for Cuda. But they will not do it. It's the best example of artificial.



Mantle was developed for GCN. It's a specific low level API for a certain architecture. It's not able to run on other hardware. That is a artificial lockdown. Dice can use DX or OpenGL and trying to convince Microsoft for a better support in DX. But they decided to create an own API for GCN. Really that is right the definition of artificial and excactly the same like Cuda.

That is not an artificial limitation, that is a natural technical limitation, that was his point.

So, why would a developer use Mantle if they still need another DX path? The reason for low level API is that you dont want to optimize for other plattforms. More APIs mean more work, more investment. Mantle makes only sense if you start to ignore other vendors. .

Yes, and before they already had to do that for PS3, Xbox360, and PC. I'm not sure why you think they won't put in the word for multiple code paths.

And there is no cross plattform development between Mantle and X1 and PS4. Both using their own software stack.

Yes there is. AMD specifically mentioned that the proliferation of GCN in the consoles is what allows this to make sense for game developers.

Agree with a lot of this. Understand why AMD is going in this direction, but long term for gamers it's harmful. Problem is, DirectX is just really inefficient (not to mention Windows) so from that perspective I welcome anything that alleviates the wasted processor cycles.

The perfect solution to me would be a close to the metal, open graphics API on Linux. And no OpenGL is not good enough, in fact it's a bit of a mess the last few years.

You've never done low level hardware development before have you?

 

[sushiwarrior]

Anytime the homogenization of console and PC games is served as a sh*t-sandwich, (especially when graphics are involved), all PC gamers should be concerned.

Not PC gamers who are more concerned about actual performance, only "PC Master Race" people who are worried about keeping their platform as closed off and inefficient as possible.

 

[AnandThenMan]

Mantle was developed for GCN. It's a specific low level API for a certain architecture. It's not able to run on other hardware. That is a artificial lockdown.

Not artificial anything, it's the reality of the hardware/software ecosystem.

 

[AnandThenMan]

You've never done low level hardware development before have you?

Yes, first processor I learned on was the 6809.

 

[OCGuy]

Not PC gamers who are more concerned about actual performance, only "PC Master Race" people who are worried about keeping their platform as closed off and inefficient as possible.

I, like most people, will watch where it goes. If it is truly increased performance, good for them. The market will sort it all out.

If something is sacrificed in order to obtain that performance, then it will fail badly.

 

[jvroig]

Mantle was developed for GCN. It's a specific low level API for a certain architecture. It's not able to run on other hardware. That is a artificial lockdown. Dice can use DX or OpenGL and trying to convince Microsoft for a better support in DX. But they decided to create an own API for GCN. Really that is right the definition of artificial and excactly the same like Cuda.

You keep on insisting on your own terminology despite the very meaning of "artificial" and how I've already tried to explain it to you.

I'm not sure if you just don't get it, or you don't want to get it. If you think that CUDA and Mantle are "artificially locked down" despite the fact that they are inherently bound to the hardware arch they are targeting ("inherently", meaning by its very nature, therefore does not require artificial means to produce), then you have a very different meaning of "artificial" compared to the rest of the world, or you are simply trolling.

Either way, I will not waste any more of my time. Believe what you wish.

 

[BD2003]

This is nothing like the previous times precisely because there are already well supported common APIs and multiple competitors in the market place. OpenGL and DirectX won't go anywhere, and Intel and nVidia will continue to chug ahead. It would be moronic to release a game which only support Mantle or even favor Mantle too much. The vast majority of computers out there don't even have a discreet graphics card and use built in Intel graphics...

How could it not favor mantle too much? Using glide gave you HUGE gains over DX on 3Dfx cards, it was night and day. Nvidia will continue to chug ahead for a time, but no gamer would seriously consider using them if mantle gains traction. Why would anyone in their right mind pay the same price for an Nvidia card that performs 30% slower than AMD in a bunch of high profile games? It isnt enough that nvidia and DX/OGL continues to exist, they have to actually remain competitive.

I hope the industry flat out rejects it.

 

[Arkadrel]

This new api, offering high boosts to performance, by codeing low to metal is cool.
I like that its not locked down by a OS, this is bound to help Linux crowd alot.

 

[Saylick]

Anytime the homogenization of console and PC games is served as a sh*t-sandwich, (especially when graphics are involved), all PC gamers should be concerned.

It sounds like you are implying that console games will look graphically similar to PC games as a result of this. I don't think this would be the case considering high-end PC hardware will be faster than console hardware.

I don't know if this analogy is the correct way to view Mantle, so please correct me if it's not valid, but the way I see things, Mantle sounds like a driver update specific to GCN that potentially promises a substantial performance boost across the board for GCN cards, with the premise that most games in the future will be optimized for GCN.

If Mantle was wrapped up and delivered as a ridiculously awesome driver update, would people's perceptions change?

 

[GaiaHunter]

Developers wouldn't port using a specific API for only AMD or only NVIDIA if they weren't using those tools already for developing for consoles.

PC exclusive games will only use DX.

Mantle won't lock any features to AMD - it will just run faster because it is optimized for AMD just like tons of software is optimized for Intel since Intel is the biggest player by far in the market.

AMD isn't the biggest player in the PC front, but when you combine PC gaming and console gaming, it will be.

PhysX (hardware accelerated version) never got any traction since NVIDIA didn't have a big enough lead vs AMD for developers to care.

Allowing AMD to have APUs in all the consoles was very short sighted from other players, allowing AMD to clearly benefit from optimizations for their hardware in the consoles to bridge into the PC.

 

[Despoiler]

Techspot is confirming that Mantle is an open API.

http://www.techspot.com/news/54134-...y-mantle-api-to-optimize-gpu-performance.html

We've been told at the GPU14 Tech Day event that the Mantle API is open, so theoretically Nvidia could purpose the technology in their GPUs. It should also make cross-development between PC and console games a lot easier, and also more incredible for those with a high-performance AMD GPU.

I think that Mantle is more of a shot at MS. DX is trash in terms of raw performance. It's ridiculously unoptimized compared to low level access.

 

[EightySix Four]

Yes, first processor I learned on was the 6809.

Then explain exactly how you would intend to build a single API that supports multiple highly different architectures at a "close to the metal" level? If you have a solution for that, the whole industry would love to hear it. Being able to write code on one platform that supports multiple architectures at a low level with no overhead would change the world.