When do you think the first DX12 games will arrive?

[soccerballtux]

Yup. I'm sure people will of course push their favorite vendor because it supports xyz direct x 12 feature but none of those features are relevant or will be relevant to your card when they arrive.

well, that's kinda what I'm asking about. Since AMD went to DX11 so early, if you have an AMD card, you're much more likely to get benefit from DX12 than NVidia

 

[bystander36]

well, that's kinda what I'm asking about. Since AMD went to DX11 so early, if you have an AMD card, you're much more likely to get benefit from DX12 than NVidia

That doesn't make sense to me. What does DX11 have to do with DX12? Given that only Nvidia supports DX 12_1 feature sets, I don't understand.

Now AMD's Mantle push, with their GCN arch might prove useful for DX12, but DX11 has little to do with DX12.

 

[soccerballtux]

That doesn't make sense to me. What does DX11 have to do with DX12? Given that only Nvidia supports DX 12_1 feature sets, I don't understand.

Now AMD's Mantle push, with their GCN arch might prove useful for DX12, but DX11 has little to do with DX12.

because many of the performance improvements for DX12 are building on top of DX11 features.

with 12_1 NVidia basically sponsored the additions

 

[bystander36]

because many of the performance improvements for DX12 are building on top of DX11 features.

with 12_1 NVidia basically sponsored the additions

Again, that doesn't make much sense. Nvidia and AMD have been on DX11 for a very long time. Being first to support DX11 doesn't mean much.

 

[RussianSensation]

The timecontext of the twitter is still holidays 2016 (or later).

That's funny considering you are one of the most vocal defenders of how DX12 was a key selling point for the entire Maxwell GPU architecture since GTX750/750Ti launched and now you are finding information that DX12 won't matter much until late 2016 which is what most people have told you for the last 1.5 years. I mean you posted so much FUD about GCN vs. Maxwell vs. Kepler vs. Fermi DX12 and so much implication of how it's a big deal and now we continue to see reality unfolding that there isn't a single true DX12 game build from the ground-up on the horizon for a while, by which point we'll have 16nm HBM2 GPUs from both companies. :sneaky:

 

[ShintaiDK]

That's funny considering you are one of the most vocal defenders of how DX12 was a key selling point for the entire Maxwell GPU architecture since GTX750/750Ti launched and now you are finding information that DX12 won't matter much until late 2016 which is what most people have told you for the last 1.5 years. I mean you posted so much FUD about GCN vs. Maxwell vs. Kepler vs. Fermi DX12 and so much implication of how it's a big deal and now we continue to see reality unfolding that there isn't a single true DX12 game build from the ground-up on the horizon for a while, by which point we'll have 16nm HBM2 GPUs from both companies. :sneaky:

Nice try on a personal attack, but again you obviously try too much. But maybe you should read the post again and Johans twitter.

 

[Pottuvoi]

overhead benefits are all anyone should really care about. The new features take so long to get used....

Also most of the new features are also coming to DX11.3, so it's not the biggest reason to go for DX12.

Biggest reason is to better performance and to have more control on you code. (not sure if driver still can mess everything.)

More on DX12.
DX12 in MSDN
DirectX 11 to DirectX 12 porting guide

 

[LFCNZ]

Sorry if it has been stated already but does anyone know what current games are getting DX12 patches?

I believe Project Cars is?

Thanks

 

[Azix]

overhead benefits are all anyone should really care about. The new features take so long to get used....

Exactly, lots of low hanging fruit there to benefit not only GCN but also nV's products, only from the overhead improvements. The rest that's enabled by DX12 will come later, as usual.

Some features might come later, at least one useful feature is coming earlier and we can be glad for that. Asynchronous shaders seems to be making its way in there for this year and early next year to add to the reduced overhead and multithreading of dx12.

The reason dx12 adoption might be so fast is really down to GCN being in the consoles. They are already messing around with some of those features and when dx12 exposes them it would be sensible to bring the same features to the PC versions. Another reason is the fact xb1 is getting dx12 and shares games with PC that ps4 doesn't have (eg. timed exclusive tomb raider could exploit those features).

And the fact dx12 just got officially released does not mean games will take long to come. Devs have been on it for months either from learning mantle for the future dx12 transfer or dx12 directly.

 

[soccerballtux]

Also most of the new features are also coming to DX11.3, so it's not the biggest reason to go for DX12.

Biggest reason is to better performance and to have more control on you code. (not sure if driver still can mess everything.)

More on DX12.
DX12 in MSDN
DirectX 11 to DirectX 12 porting guide

oh, interesting. is that NOT coming to 11.3?

 

[Elixer]

Also most of the new features are also coming to DX11.3, so it's not the biggest reason to go for DX12.

Biggest reason is to better performance and to have more control on you code. (not sure if driver still can mess everything.)

What? DX12's new features will not be backported to DX11.x, so, why do you think that 'most of the new features' are going to make it into a DX 11 update?
*edit, unless you mean only for WDDM 2

 

[antihelten]

What? DX12's new features will not be backported to DX11.x, so, why do you think that 'most of the new features' are going to make it into a DX 11 update?

Maybe because D3D 11.3 can support feature level 12_1 (plus a bunch of the optional features beyond feature level 12_1).

In other words, feature level 12_1 is not exclusive to D3D 12.

Of course Pottuvoi might have been referring to features other than those that are part of the various feature levels, in which case he would be wrong, but when people talk about DX12 features, they tend to refer to the feature levels.

 

[Spanners]

Nice try on a personal attack, but again you obviously try too much. But maybe you should read the post again and Johans twitter.

Nice try at a rebuttal but the thing about personal attacks is, and this shouldn't be surprising it's in the name, people have to attack you personally for them to have occurred. Saying he thinks you're wrong or have made a poorly reasoned/inconsistent argument for the nth time is not a personal attack. I think you need to try harder.

 

[Elixer]

Maybe because D3D 11.3 can support feature level 12_1 (plus a bunch of the optional features beyond feature level 12_1).

In other words, feature level 12_1 is not exclusive to D3D 12.

Of course Pottuvoi might have been referring to features other than those that are part of the various feature levels, in which case he would be wrong, but when people talk about DX12 features, they tend to refer to the feature levels.

I think we can agree that those feature levels still requires WDDM 2.0, and won't work on WDDM 1.x.

 

[antihelten]

I think we can agree that those feature levels still requires WDDM 2.0, and won't work on WDDM 1.x.

yes, so? D3D 11.3 works with WDDM 2.0

 

[ViRGE]

I think we can agree that those feature levels still requires WDDM 2.0, and won't work on WDDM 1.x.

Newer feature levels will work on DX11 via 11.3. That's more or less the whole purpose for 11.3: to expose those new features to the high level API, since DX12's low level API isn't going to be as easy to use for non-guru programmers.

http://www.anandtech.com/show/8544/microsoft-details-direct3d-113-12-new-features

Besides the low level API itself, technically there's a feature or two that can't be brought over to DX11, though I don't know if there's a good list at this time.

 

[ThatBuzzkiller]

Newer feature levels will work on DX11 via 11.3. That's more or less the whole purpose for 11.3: to expose those new features to the high level API, since DX12's low level API isn't going to be as easy to use for non-guru programmers.

http://www.anandtech.com/show/8544/microsoft-details-direct3d-113-12-new-features

Besides the low level API itself, technically there's a feature or two that can't be brought over to DX11, though I don't know if there's a good list at this time.

Those two major features are ExecuteIndirect and Asynchronous Compute ...

 

[ibex333]

Even when the first DX12 games arrive it wont be a "super shocking" improvement over DX11. Maybe some lower end cards will benefit from faster FPS but it's all very relative to the system in question, parts used, etc, etc. For people playing at lower resolutions with lower draw distance it wont even matter at all. But money wise I am sure the difference will be very, very significant.

 

[ShintaiDK]

Even when the first DX12 games arrive it wont be a "super shocking" improvement over DX11. Maybe some lower end cards will benefit from faster FPS but it's all very relative to the system in question, parts used, etc, etc. For people playing at lower resolutions with lower draw distance it wont even matter at all. But money wise I am sure the difference will be very, very significant.

The GPU part wont really be faster. The main benefits is on the CPU part.

 

[Despoiler]

Fable Legends Win10 beta is coming this month.

http://wccftech.com/dx12-powered-fable-legends-screens-windows-10-beta-month/

 

[NTMBK]

The GPU part wont really be faster. The main benefits is on the CPU part.

Not quite; async shaders and memory uploads let you get better utilisation from your GPU.

 

[bystander36]

Not quite; async shaders and memory uploads let you get better utilisation from your GPU.

Only if utilized. At first, you can expect very little change on the GPU side, until a bit later. They are going to want to wait until more GPU hardware can take advantage of those techs before game engines around it.