DirectX and OpenGL to get low level access according to leaked statements (GDC 2014)

[Carfax83]

I'm sure this thread will eventually get locked, but like I and many others have been saying all a long, Microsoft has no intentions of surrendering their PC gaming monopoly (even though they've taken it for granted over the years) and will not allow Mantle to supersede DirectX.

Microsoft has more experience with this sort of thing than both AMD and NVidia combined.. I'm sure they can pull it off, while retaining their broad architectural compatibility.

DirectX & OpenGL catching up with Mantle, will offer low-level access to reduce draw overhead..

“In this session we will discuss future improvements in Direct3D that will allow developers an unprecedented level of hardware control and reduced CPU rendering overhead across a broad ecosystem of hardware.”

"Driver overhead has been a frustrating reality for game developers for the entire life of the PC game industry. On desktop systems, driver overhead can decrease frame rate, while on mobile devices driver overhead is more insidious–robbing both battery life and frame rate. In this unprecedented sponsored session, Graham Sellers (AMD), Tim Foley (Intel), Cass Everitt (NVIDIA) and John McDonald (NVIDIA) will present high-level concepts available in today’s OpenGL implementations that radically reduce driver overhead–by up to 10x or more."

 

[3DVagabond]

YAY COMPETITION!!! :thumbsup:

Why is it so many people kick and scream anytime their favorite company gets a little competition? What do all these people who claim that their CPU's aren't bottle necked and draw calls aren't an issue have to say now? This is good for all of us.

 

[Carfax83]

YAY COMPETITION!!! :thumbsup:

Why is it so many people kick and scream anytime their favorite company gets a little competition? What do all these people who claim that their CPU's aren't bottle necked and draw calls aren't an issue have to say now? This is good for all of us.

Well I can easily answer this. Competition is good, and reducing CPU bottlenecks and increasing draw calls is great.

But there's a right way and a wrong way to go about doing that. What we're seeing right now from Microsoft and the Khronos Group is the right way.

Mantle is the wrong way....imo.

 

[hungtran]

This is a win for AMD as it makes them finally seem like innovators again who were ahead of the curve. And wide spread reduction of CPU rendering overhead should help them sell more APUs.

 

[sontin]

This is a win for AMD as it makes them finally seem like innovators again who were ahead of the curve. And wide spread reduction of CPU rendering overhead should help them sell more APUs.

What?!
OpenGL 4.4 was published in july 2013 and nVidia has supported it since then.

Mantle was announced end of september...

There has been a push for less driver overhead from every party without the introduction of new APIs.

 

[VulgarDisplay]

Well I can easily answer this. Competition is good, and reducing CPU bottlenecks and increasing draw calls is great.

But there's a right way and a wrong way to go about doing that. What we're seeing right now from Microsoft and the Khronos Group is the right way.

Mantle is the wrong way....imo.

You're under the assumption that AMD did it the wrong way because they were using Direct 3d HLSL obviously with microsoft's blessing?

I wouldn't be surprised if it comes out later that this new direct x uses mantle for AMD cards.

 

[DarkKnightDude]

From Intel, AMD and Nvidia?

I guess they all agree on one thing.

 

[hungtran]

Intel stands to lose the most if bottleneck for gaming moves from the CPU to GPU. It's interesting they would agree here.

 

[desprado]

Microsoft has more experience with this sort of thing than both AMD and NVidia combined.. I'm sure they can pull it off, while retaining their broad architectural compatibility.

DirectX & OpenGL catching up with Mantle, will offer low-level access to reduce draw overhead..

That comment.Agreed yes even if u combine intel,Nvidia and AMD than still MS experience will win and it is better that now they are making a low level API so everyone can be happy with no additional Patch required.

 

[Will Robinson]

YAY COMPETITION!!! :thumbsup:

Why is it so many people kick and scream anytime their favorite company gets a little competition? What do all these people who claim that their CPU's aren't bottle necked and draw calls aren't an issue have to say now? This is good for all of us.

NV is involved...now it's important...:thumbsup:
Just like cryptomining's sudden importance.

 

[BFG10K]

This is one of the reasons why Mantle will never take off.

Intel stands to lose the most if bottleneck for gaming moves from the CPU to GPU. It's interesting they would agree here.

The extra performance would still help Intel's CPUs though, and put them in a better light. Most games are already GPU bottlenecked anyway if you run settings expected for a given level of performance (i.e. nobody buys a Titan to play @ 1080p without AA).

I'm actually more surprised to see an AMD guy there given it completely undermines their whole Mantle strategy.

 

[hungtran]

Actually, benchmarks will all start looking like the current i7+290x mantle vs fx8350 + 290x mantle. The AMD CPUs will get much more boost. If most games will be GPU bottlenecked, then this whole exercise with DirectX reducing CPU rendering overhead will be pointless.

 

[Carfax83]

You're under the assumption that AMD did it the wrong way because they were using Direct 3d HLSL obviously with microsoft's blessing?

I wouldn't be surprised if it comes out later that this new direct x uses mantle for AMD cards.

We already went over this in the old thread. The best case scenario for Mantle is for it to become subsumed into a future Direct3D version..

It looks like Microsoft is going the extension route though, so I don't know how it's going to compare to Mantle.

 

[3DVagabond]

This is one of the reasons why Mantle will never take off.


The extra performance would still help Intel's CPUs though, and put them in a better light. Most games are already GPU bottlenecked anyway if you run settings expected for a given level of performance (i.e. nobody buys a Titan to play @ 1080p without AA).

I'm actually more surprised to see an AMD guy there given it completely undermines their whole Mantle strategy.

I think most games are GPU bottlenecked because of certain settings that incur incredible overhead. Especially for the IQ improvement they give, or lack there of.

I also believe that most AAA games are also CPU limited. If not by a pure bottleneck then at least by the relatively limited CPU resources that have been available. There are a lot of tricks optimizations that are done purely to limit cpu overhead that effect the artists. Add to that GPU performance is advancing far faster than cpu performance and I can see the issue only getting worse, and quickly.

As far as AMD's strategy... Who knows what it is exactly? They did say they had been in communication with Microsoft prior to going ahead with Mantle. The quote, if I can paraphrase it, was along the lines of, "We didn't blindside Microsoft with Mantle. Our relationship with them was and still is very important and we wouldn't want to jeopardize that." More or less. They also said that they hoped others would adopt Mantle. While it's too early to tell exactly where this is going and what they are all up to, none of this contradicts that and AMD is well known for simply offering their work up for open standards.

 

[Carfax83]

Actually, benchmarks will all start looking like the current i7+290x mantle vs fx8350 + 290x mantle. The AMD CPUs will get much more boost. If most games will be GPU bottlenecked, then this whole exercise with DirectX reducing CPU rendering overhead will be pointless.

Why is it pointless? Developers will just find ways to use the extra CPU power that's available.

Longer draw distances, more detailed environments, more sophisticated physics and A.I..

They can do lots of things with the additional CPU power..

 

[kukreknecmi]

You got it wrong, so did DSO. It's about buffer_storage, bindless , storage buffer object (and some indirect multi draw).

For Win/Dx side, there's nothing noted yet. They've inspired from bindless and implementation is mostly about sparse tex / bindless (tiled resourced kinda thing) for now.

If someone is expecting post-July-September validated gl extensions at 2013 to be a game changer for say at least 1 year ahead of time, nice dream.

Unless you have Win/Dx level support / backers with some well documented development tools, it only happens far far away.

In case of reducing driver interaction around %75 even, you WILL need kinda Sellers / Everitt / McDonald's level developers around which i found it hard to believe and encourage focusing on GL more.

Side note : Nvidia's bindless and ARB's are different, even so AMD's implementation.

 

[3DVagabond]

NV is involved...now it's important...:thumbsup:
Just like cryptomining's sudden importance.

So much for the fear that Mantle is going to fracture the community with all of the hardware manufacturers creating separate API's, buying different GPU's for different games, etc.

 

[OCGuy]

And that is why Mantle was not a wise investment for AMD.

 

[3DVagabond]

And that is why Mantle was not a wise investment for AMD.

Did you miss the part where AMD is part of the group? :\

 

[OCGuy]

NV is involved...now it's important...:thumbsup:
Just like cryptomining's sudden importance.

All players benefit, which is what is necessary for increased natural developer adoption. If Physx cant take off with their marketshare, anything even more niche will have a more difficult battle.

 

[SiliconWars]

Well I can easily answer this. Competition is good, and reducing CPU bottlenecks and increasing draw calls is great.

But there's a right way and a wrong way to go about doing that. What we're seeing right now from Microsoft and the Khronos Group is the right way.

Mantle is the wrong way....imo.

And what do you think has given Microsoft and Khronos the hurry up? This is happening BECAUSE of Mantle.

We already went over this in the old thread. The best case scenario for Mantle is for it to become subsumed into a future Direct3D version..

It looks like Microsoft is going the extension route though, so I don't know how it's going to compare to Mantle.

You know what they say about putting lipstick on a pig.

 

[Spjut]

I just hope that the improvements in the next DirectX and OpenGL will benefit today's DX11.x cards as well, like how DX10 cards supported some of the new features in DX11.
My selfish reasons aside, I think that would help getting studios supporting the new versions.

But on the DirectX side, most studios will still focus on DX11.0 due to Windows 7. For all the complaints about Mantle being AMD only, it's a huge benefit that it works on Windows 7.

 

[GaiaHunter]

It is always interesting that something that couldn't possibly succeed makes others react to it.

Maybe it is not success wasn't actually guaranteed.

 

[desprado]

Some people think this is happening because of Mantle but it is fact that If DirectX get a low access than it is sure the end of mantle and no one will be willing to spend extra time,extra man power and money unless AMD pays for it.

 

[desprado]

I just hope that the improvements in the next DirectX and OpenGL will benefit today's DX11.x cards as well, like how DX10 cards supported some of the new features in DX11.
My selfish reasons aside, I think that would help getting studios supporting the new versions.

But on the DirectX side, most studios will still focus on DX11.0 due to Windows 7. For all the complaints about Mantle being AMD only, it's a huge benefit that it works on Windows 7.

yes i think they will improve DX 11 so win 7 can advantage.