DirectX 12

[Despoiler]

When there is a need for something, everything happens faster. Even DX11 did not take long to be used. DX11 was there the moment the first DX capable cards were sold (AMD 5000 cards), and the first game came soon after with Dirt 2.

With Mantle exposing the problems of DX, I'm sure it won't take all that long. DX9 was also used quite quickly. DX10, on the other hand, took a long time and never did become widespread.

We'll have to wait and see about Dx12.

Sure, the DX11 feature set was finalized with Nvidia, AMD, and game devs well before the functional parts were released. That is normal with a standards process. Everyone has to be on board so they can design the chips and games for a target launch. No one, but MS knows anything about DX12, which alludes to the very beginning of development. At best they started when Mantle was announced. If that turns out to be the case, and it's a very good bet, we are looking 1-2 years from now for release and early adoption. Remember that AMD has been working on Mantle for 2-3 years and they didn't worry about backwards compatibility.

 

[bystander36]

Sure, the DX11 feature set was finalized with Nvidia, AMD, and game devs well before the functional parts were released. That is normal with a standards process. Everyone has to be on board so they can design the chips and games for a target launch. No one, but MS knows anything about DX12, which alludes to the very beginning of development. At best they started when Mantle was announced. If that turns out to be the case, and it's a very good bet, we are looking 1-2 years from now for release and early adoption. Remember that AMD has been working on Mantle for 2-3 years and they didn't worry about backwards compatibility.

While I'm sure DX12 is in development still, if it is announced, it probably isn't the beginning of development either. Some of these DX versions are quicker than others to be finalized as well.

I'm just saying that the years and years that CakeMonster was stating is a worst case scenario. There have been quick releases and long ones.

 

[96Firebird]

This is a good point. Why even bother coding for DX12 if it's a) only supported in some tiny number of super high end PCs or worse yet possibly b) is only available for windows 9

If we change a few words around, we can ask the same question about another API... :sneaky:

 

[VulgarDisplay]

I for one think amd and nvidia have known about dx 12 for a while and that amd's mantle driver will basically work for dx12.

 

[VulgarDisplay]

If we change a few words around, we can ask the same question about another API... :sneaky:

Try so we can tell you how wrong you are. Need to change entire sentences maybe.

 

[Adampa1006]

Remember how long it took to get dx 11 native games? Games that actually looked like a true step above dx 9. Even some dx 9 games like Witcher 2 can keep up in today's market. I'm an enthusiast and want dx 12 native games here ASAP as I will spend the money to buy the supporting hardware, but let's be realistic that's years away. I'll watch and see though.

 

[Despoiler]

While I'm sure DX12 is in development still, if it is announced, it probably isn't the beginning of development either. Some of these DX versions are quicker than others to be finalized as well.

I'm just saying that the years and years that CakeMonster was stating is a worst case scenario. There have been quick releases and long ones.

In order for DX to compete with Mantle they need to solve all of the inherent problems in DX. The big ones being poor CPU scaling, API overhead, and driver overhead. DX has the performance properties it has because its core design is from decades ago. Do you honestly think that will be a quick iteration of DX?

 

[bystander36]

In order for DX to compete with Mantle they need to solve all of the inherent problems in DX. The big ones being poor CPU scaling, API overhead, and driver overhead. DX has the performance properties it has because its core design is from decades ago. Do you honestly think that will be a quick iteration of DX?

Do you honestly think MS hasn't been working on this for a while now? As soon as they release any software, they are already working on the sequel. Why would this be any different?

And do you have any inside info on how difficult this would be? We'd love to hear it, if you do.

Who is to say that MS hasn't had lower level code ready to go for years, but chose not release it for compatibility issues. Until further info, you cannot assume anything.

 

[Despoiler]

Do you honestly think MS hasn't been working on this for a while now? As soon as they release any software, they are already working on the sequel. Why would this be any different?

And do you have any inside info on how difficult this would be? We'd love to hear it, if you do.

Who is to say that MS hasn't had lower level code ready to go for years, but chose not for compatibility issues. Until further info, you cannot assume anything.

No one, but MS knows anything about DX12. MS would have to involve AMD, Intel, and Nvidia in order to have hardware support. Industry folks have commented that MS's interest in DX12 is recent. Therefore MS hasn't been working on it beyond when Mantle was announced.

I do have knowledge, but it doesn't need to be inside to be relevant. It took AMD 2-3 years to develop Mantle to the point it is right now. That's without backwards compatibility nor having to tie the dang thing into the OS like MS does.

 

[bystander36]

No one, but MS knows anything about DX12. MS would have to involve AMD, Intel, and Nvidia in order to have hardware support. Industry folks have commented that MS's interest in DX12 is recent. Therefore MS hasn't been working on it beyond when Mantle was announced.

I do have knowledge, but it doesn't need to be inside to be relevant. It took AMD 2-3 years to develop Mantle to the point it is right now. That's without backwards compatibility nor having to tie the dang thing into the OS like MS does.

We'll have to wait and see. And it can be to their advantage to have the backwards compatibility in terms of speed of development. That means they have lots of code to reuse. They are not starting from scratch. They do not have to rewrite the whole API for major improvements, not that they won't do that either.

DX could be majorly improved simply by adding good multithreading. Though how difficult it would be is not known to us. Adding the lower level extensions that exist now, into the heart of DX would also be a quick easy improvement.

We also do not know if there is going to be any new hardware requirements for DX12. It may simply be software changes.

 

[PliotronX]

Oh lord, the pessimist in me comes out when I see the Qualcomm logo..

 

[Gloomy]

Oh lord, the pessimist in me comes out when I see the Qualcomm logo..

DirectX 12 is Born Mobile. Are you?

 

[DarkKnightDude]

DirectX 12, exclusive to Windows 9 and Windows phone!

 

[TechFan1]

Some tweets about DX12 by Francois Piednoel that seemed interesting.
-There are new techs that force re-architecture,Shared memory between CPU and GPU, MS has been working on it for long time
-Up to 2014 no breakthrough techs ,only incremental for 5 years. When your GPU can use pointers from CPU, it change the game
-Rightnow,GPU world is about throughput and hiding latencies with massive threading, new CPU<=>GPU latencies are going to change the game(s).
-[Microsoft] did not ignore, they were waiting that all the planets get aligned ...

 

[Despoiler]

We'll have to wait and see. And it can be to their advantage to have the backwards compatibility in terms of speed of development. That means they have lots of code to reuse. They are not starting from scratch. They do not have to rewrite the whole API for major improvements, not that they won't do that either.

DX could be majorly improved simply by adding good multithreading. Though how difficult it would be is not known to us. Adding the lower level extensions that exist now, into the heart of DX would also be a quick easy improvement.

We also do not know if there is going to be any new hardware requirements for DX12. It may simply be software changes.

Weren't you the one asking me about insider knowledge on how I knew so much about all of this? Yet, here you are making claims you can just fairy dust and duct tape DX11 into DX12 that can match Mantle. Let me say this again, DX's problems are because the core architecture was built for a hardware landscape that has long since past. You don't just bandaid an out of date architecture to support a modern landscape. You have to design and build an architecture that suits your needs from the ground up. There is no quick and easy. Especially not in Microsoft land where the amount of garbage they have between your hardware is part of the problem.

Some tweets about DX12 by Francois Piednoel that seemed interesting.
-There are new techs that force re-architecture,Shared memory between CPU and GPU, MS has been working on it for long time
-Up to 2014 no breakthrough techs ,only incremental for 5 years. When your GPU can use pointers from CPU, it change the game
-Rightnow,GPU world is about throughput and hiding latencies with massive threading, new CPU<=>GPU latencies are going to change the game(s).
-[Microsoft] did not ignore, they were waiting that all the planets get aligned ...

Oh boy an Intel dev. I'd be talking mad smack too if my competitor figured out a way to make their inferior CPUs not relevant in the gaming performance debate. It doesn't even matter if Mantle sticks around because AMD won by forcing the elimination of the CPU bottleneck in gaming.

Warning issued for inflammatory language.
-- stahlhart

 

[R0H1T]

Some tweets about DX12 by Francois Piednoel that seemed interesting.
-There are new techs that force re-architecture,Shared memory between CPU and GPU, MS has been working on it for long time

Hmmm Interesting seems like he's talking about HSA/HUMA

-Up to 2014 no breakthrough techs ,only incremental for 5 years. When your GPU can use pointers from CPU, it change the game

Sounds like Intel since 2010 D:

-Rightnow,GPU world is about throughput and hiding latencies with massive threading, new CPU<=>GPU latencies are going to change the game(s).

Considering how bad Intel's graphics capabilities have been over the years(still are) no wonder they've been praying at MS' altar but the fact that AMD/Mantle has helped them is kinda ironic, again not for the first time !

-[Microsoft] did not ignore, they were waiting that all the planets get aligned ...

Oh but they had to, & were forced into it by the dwindling sales of Windows & Pcs in general, otherwise it'd have taken another decade for them to get this through to the market, not some planetary alignment

 

[bystander36]

Weren't you the one asking me about insider knowledge on how I knew so much about all of this? Yet, here you are making claims you can just fairy dust and duct tape DX11 into DX12 that can match Mantle. Let me say this again, DX's problems are because the core architecture was built for a hardware landscape that has long since past. You don't just bandaid an out of date architecture to support a modern landscape. You have to design and build an architecture that suits your needs from the ground up. There is no quick and easy. Especially not in Microsoft land where the amount of garbage they have between your hardware is part of the problem.

I said we don't know how long it will take. We don't know where we are and we don't know that it'll take years and years and years to be delivered. You don't know, I don't know. MS probably has an idea, however. I simply am stating that you don't know crap, and I am not claiming to know any more than you. Just that you cannot say that it will take X years before we see it.

Give it a rest.

 

[imaheadcase]

Fact checking implies that you check the premise, otherwise said fact is merely contention.

In your case the premise is that AMD is losing significant gaming market share due to mining (fun fact: your premise also has a premise, meaning you need to back up the first premise, AND connect it to the subsequent premise, in order to impart validity to your overall claim).

So yes, relevant sources are sill needed

Are you serious? You don't need to fact check that, only 2 games have mantle support right now, and one of them BF4 is "iffy" it gets broke every patch, and still bad performance on some series cards. AMD is not gunning towards the gaming market, never has. Gaming market has marginal impact on sales vs OEM/retail sales.

 

[frozentundra123456]

Are you serious? You don't need to fact check that, only 2 games have mantle support right now, and one of them BF4 is "iffy" it gets broke every patch, and still bad performance on some series cards. AMD is not gunning towards the gaming market, never has. Gaming market has marginal impact on sales vs OEM/retail sales.

Actually it is only BF4. Thief is waiting for a patch that is supposed to come soon. If BF 4 is any indicator, could be a long time before it is working properly on all gcn cards.

 

[RampantAndroid]

Actually it is only BF4. Thief is waiting for a patch that is supposed to come soon. If BF 4 is any indicator, could be a long time before it is working properly on all gcn cards.

And one wonders if anyone will care, given the lukewarm reception Thief has had thus far.

 

[frozentundra123456]

And one wonders if anyone will care, given the lukewarm reception Thief has had thus far.

Exactly, and especially since an i3 is more than adequate cpu wise. Really a strange choice of a game for mantle, IMO.

 

[PliotronX]

DirectX 12 is Born Mobile. Are you?

Ehhh, something just doesn't seem right because normally when games are cross platform in any way they are coded for the lowest common denominator.

 

[Black Octagon]

Are you serious? You don't need to fact check that, only 2 games have mantle support right now, and one of them BF4 is "iffy" it gets broke every patch, and still bad performance on some series cards. AMD is not gunning towards the gaming market, never has. Gaming market has marginal impact on sales vs OEM/retail sales.

Yes I am. Your post stated that the mining craze was resulting in decreased gaming market share for AMD and that this decreased market share would detract developers away from Mantle.

This series of premises (not backed up by sources) was offered to argue that Mantle uptake may not be so great.

Instead of defending your (questionable) premises and the overall logical validity of your argument, you simply pointed to the fact that Mantle uptake thus far (in these very early days for the new API) has been small. It's a different premise to what you originally offered, and a questionable one at that because:
1) Developers were the ones who requested Mantle,
2) BF4 was always presented as the first test case for Mantle,
3) Where matured, Mantle can be applied to whole game ENGINES and not just individual games, and
4) It's REALLY early days for Mantle.

To me, based on your kind of thinking, we should dismiss all innovation purely because it doesn't immediately achieve significant market penetration.

 

[ShintaiDK]

1) Developers were the ones who requested Mantle,
2) BF4 was always presented as the first test case for Mantle,
3) Where matured, Mantle can be applied to whole game ENGINES and not just individual games, and
4) It's REALLY early days for Mantle.

1: If developers asked for Mantle. Why is everyone avoiding it then, unless they are heavily sponsored?

http://www.amd.com/us/products/technologies/mantle/Pages/mantle.aspx#2

2, 3 and 4:
Mantle is a stillborn API. It only works on GCN with 35% dGPU and 18% CPU share, both declining, specially the last one. Consoles rejected it as well. Also the Frostbite3 engine was to have Mantle build in. Yet no other Frostbite3 games besides BF4 seesm to be getting Mantle at all now.

The future is DirectX and OpenGL. Mantle is no part of that future.

A year ago AMD said DX was dead and there would be no DX12. And if GCN1.1 is the competition to Maxwell. Its clear what happend. AMD ran out of R&D funding.

 

[7stars]

to me if devs (sponsored or not) had not asked for something more...
maybe DirectX 12 would never have seen the light
someone who asks an interesting question:
http://monimega.com/blog/2014/03/07/microsoft-the-company-which-is-always-in-need-of-aid-amd-mantle-directx-12/