DirectX 12: The future of it within the console gaming space (specifically the XB1)

How about Mantle being derived from the consoles' low level apis and MS simply taking their console's lower level version of DX, intergrating a nVidia path and porting it to a PC.
Testing alone could drag a simple port on for ages. The console API is far removed from the PC API because it targets one hardware configuration. The console API basically has the same frontend for developers to use, but a completely different backend (which is why it performs so differently). A simple port strikes me as unreasonable given the complexity of hardware configurations it would have to deal with, though I may be wrong.

We aren't talking about DX being released to the public or near completion by GDC on March 20th. For all we know DX12 may have a 2015 release date along side Windows 9.
Going by previous DX updates, we're due a new DX. DX11 launched 2009; it's four years old. DX9 lasted 4 years before being replaced by DX10. DX10 lasted 3 years before its replacement. So unless the DX team have been swanning around doing very little, chances are their vision for DX12 has been in progress for some time and is fairly close to release (one would expect this year).
 
We aren't talking about DX being released to the public or near completion by GDC on March 20th. For all we know DX12 may have a 2015 release date along side Windows 9.

Exactly. We know nothing about DX 12, except that it will exist.
But my main point isnt if it is a response to Mantle or not. That´s mere semantic. My doubt is: will it require new hardware? And as my previous post seems to point, i do not believe that will happen!
 
Part of the process for developing DX is significant and long-term communication between Microsoft and the major stakeholders in the graphics market, such as devs and the IHVs. One of those parties should have been AMD.

Why would AMD launch Mantle if it was part of the multi-year planning process for a new and near/mid future DX12?
Would it be trying to tilt things in favor of a specific final design direction?
 
Why would AMD launch Mantle if it was part of the multi-year planning process for a new and near/mid future DX12?
Would it be trying to tilt things in favor of a specific final design direction?

Perhaps it's simply foder for their DevRel Program or it's their attempt to catch up to Nvidia with their bindless OpenGL extensions or their CUDA program?
 
Part of the process for developing DX is significant and long-term communication between Microsoft and the major stakeholders in the graphics market, such as devs and the IHVs. One of those parties should have been AMD.

Why would AMD launch Mantle if it was part of the multi-year planning process for a new and near/mid future DX12?
Would it be trying to tilt things in favor of a specific final design direction?


My personal opinion is that Mantle grew out of an attempt by AMD to up their developer relations game where NV has traditionally enjoyed an advantage. My guess is they went to EA and said 'Work with us', EA said 'Why?' and they reply 'we have an awesome API that you can help define to make EA games look the bestest and leverage all that work you do on consoles w/PC'. As far as I'm aware Mantle is only working w/Frostbite today, DX12 is likely to be a much broader API than Mantle and will of course be vendor agnostic (are we finally getting an audio rewrite MS, I miss A3D D: ).

DX12 may well have been on the go for a while but MS wasn't ready to comment until recently or they felt that they needed to get more return from all that engineering time on XB1 APIs and the interest in Mantle made that worthwhile.
 
<conspiracy theory>

I'd think that AMD would know if there was an upcoming DX12 given they had to build the APU for the XBone. I think there will be no new hardware requirements, but a new thin API that allows more direct access and tools for both PC and Xbox.

I actually have a suspicion that Mantle (or Mantle-alike) may be integrated into DX12 as the thin API model. An accelerated way for AMD to submit Mantle to MS for DX inclusion. Maybe MS have known about Mantle for a lot longer than we think given how long MS and AMD have been working together on the XBone.

AMD does all the work, and MS get the benefit for their consoles, which they care about a lot more than PC gaming. AMD gets a market advantage over competitors in the PC space. MS don't care about Nvidia as they have no impact on the XBone. MS have taken such unilateral action before (such as dumping the hardware sound stack from DX). Nvidia will just have to suck it up and join in if they want to make DX12 cards.

Maybe it's a step too far, but I've noted that all the GDC presentations for DX12 are MS or AMD presented. All the ones listed by Nvidia are DX11.

</conspiracy theory>
 
As far as I'm aware Mantle is only working w/Frostbite today, DX12 is likely to be a much broader API than Mantle and will of course be vendor agnostic (are we finally getting an audio rewrite MS, I miss A3D D: ).

Oxide have their own Nitrous engine, and Thief's upcoming patch will be for Unreal Engine 3. There may be more announcements at GDC regarding new Mantle supporting devs.
 
You folks are bsc / delusional if you think DX12 is a direct response triggered by Mantle. Things don't move that fast. Mantle may have accelerated DX12 schedule, but it in no way is the sole and entire trigger for it.

But BRIT, if DX 12 was already on the works, and the Xbox One supported it, why didn´t microsoft anounced it as a sales weapon?

Even if no new hardware was required this would leave a lot of people thinking when choosing a new console, since only Microsoft would knew that.

Do you have any reason to believe new hardware will be required, or that Dx 12 is already advanced enough to be released soon? And i´m not asking if you hear any confidential info, but just based on whats public knowledge.
 
You folks are bsc / delusional if you think DX12 is a direct response triggered by Mantle. Things don't move that fast. Mantle may have accelerated DX12 schedule, but it in no way is the sole and entire trigger for it.

Then, maybe there are more things incluided?
 
The major benefits of lower overhead draw-calls in Mantle is better frame times in *CPU limited* situations (with less of a performance benefit when solely GPU limited).

So if DX12 is just a lower overhead API (than DX 11.X), then DX12 is *more* likely to help a CPU limited game at 30 FPS game hit 60 FPS, rather than GPU limited game at 720p game hit 1080p on Xbox One ;)

sigh, none of the games today are GPU limited. They are dev limited. The xbox gpu has way power than needed to run at 1080p. Devs just make other choices.

On the other hand, due to diminishing returns, going from 720p to 900p gets you about 90-95% of the value of going to 1080p while leaving lots of resources for better effects.
 
If you think it's going to be some miracle improvement to XB1 performance, you're equally as delusional, IMO.
Don't make the mistake of judging a console by the quality of the launch window games. Let's have this discussion again in two years when we have games that are actually using engines designed for next gen hardware instead of ported last gen engines adapted for next gen.

We haven't seen anything truly next gen yet...
 
Don't make the mistake of judging a console by the quality of the launch window games. Let's have this discussion again in two years when we have games that are actually using engines designed for next gen hardware instead of ported last gen engines adapted for next gen.

We haven't seen anything truly next gen yet...

That's not his argument. His argument is that if you think DX 12 on Windows PCs is going to magically unlock massive performance gains on a console, you are going to be disappointed.

He's not saying that 3 years down the road games on these consoles will show us much more.
 
Don't make the mistake of judging a console by the quality of the launch window games. Let's have this discussion again in two years when we have games that are actually using engines designed for next gen hardware instead of ported last gen engines adapted for next gen.

We haven't seen anything truly next gen yet...

Trials Fusion. Am I right?
 
Trials Fusion. Am I right?
Trials Fusion is a cross generation game (released for Xbox 360 as well as next gen consoles). So it uses a traditional rendering engine (just like everything else released so far).

I was talking about games that will be released 2+ years from now, made by teams that have already experience from at least one full project with the new hardware. These games no longer need to support last generation hardware (or DirectX 9/10 PCs). This is the point when gamers start seeing completely new rendering techniques and engines. Compare any PS3 launch title to TLoU. That's the difference I expect to see.
 
His argument is that if you think DX 12 on Windows PCs is going to magically unlock massive performance gains on a console, you are going to be disappointed.
Obviously. But DX12 on a console can improve performance on the console. What people forget again and again is that with DX improvements come with DX graphics kernel changes that usually do impact performance. And if XO has graphics stack similar to the Windows one (and my guess would be that it is the case in this gen more so than in the previous one) there's a lot you can gain by changing how stuff works. It's not going to be OMGZBBQ amazing gain and depending on what changes (and how) it will be bound to certain types of workload (as in: not everything is gonna get faster). But some things could and if MS talks to the developers and gathers feedback on what's sluggish, the right things could improve.

Or it could get slower and dumber. But generic statements like "changes to DX don't impact performance on existing hardware" is just weird and easily contradicted by the perf gains developers report by using vendor extensions in OGL.
 
Obviously. But DX12 on a console can improve performance on the console. What people forget again and again is that with DX improvements come with DX graphics kernel changes that usually do impact performance. And if XO has graphics stack similar to the Windows one (and my guess would be that it is the case in this gen more so than in the previous one) there's a lot you can gain by changing how stuff works. It's not going to be OMGZBBQ amazing gain and depending on what changes (and how) it will be bound to certain types of workload (as in: not everything is gonna get faster). But some things could and if MS talks to the developers and gathers feedback on what's sluggish, the right things could improve.

Or it could get slower and dumber. But generic statements like "changes to DX don't impact performance on existing hardware" is just weird and easily contradicted by the perf gains developers report by using vendor extensions in OGL.

Agreed. I've experienced first hand a performance increase for BF4 moving from Windows 7 to Windows 8. It made BF4 playable for my aging Geforce 660 GTX Phenom II X4 system

Some data points that are truthfully all over the place:

http://forums.anandtech.com/showthread.php?t=2346906

http://www.hardocp.com/article/2013...dows_7_vs_81_performance_review/#.UyMw7vldVcw

http://www.pcgameshardware.de/Battl...d-4-Windows-81-Windows-7-Performance-1108811/
 
Don't make the mistake of judging a console by the quality of the launch window games. Let's have this discussion again in two years when we have games that are actually using engines designed for next gen hardware instead of ported last gen engines adapted for next gen.

We haven't seen anything truly next gen yet...
Only Ryse comes close to be what you could expect from next gen.

Of course, NBA 2k14, FIFA 14, are 1080p and look amazing, but next gen shouldn't be about bumping the resolution alone, but better sound too.

On a different note, this guy explains in great detail why he thinks Xbox One is DirectX 12 compatible. The console could have a secret sauce.

 
Trials Fusion is a cross generation game (released for Xbox 360 as well as next gen consoles). So it uses a traditional rendering engine (just like everything else released so far).

I was talking about games that will be released 2+ years from now, made by teams that have already experience from at least one full project with the new hardware. These games no longer need to support last generation hardware (or DirectX 9/10 PCs). This is the point when gamers start seeing completely new rendering techniques and engines. Compare any PS3 launch title to TLoU. That's the difference I expect to see.

I was trying to be a kiss ass and you ruined it. I wasn't expecting the honesty.
 
Back
Top