DirectX® 12 for Enthusiasts: Explicit Multiadapter

+1 internet point for Factum.

Some people would try to have others believe that DirectX 12 wasn't even going to exist if it hadn't been for AMD showing everyone the way on like three games that supported it for a small number of users.
 
Factum said:
Care to tell me when Intel got access to Mantle? :)
Click to expand...

Intel is part of the Khronos group

https://www.youtube.com/watch?v=qKbtrVEhaw8

They got Mantle in an early stage acknowledged in this video.

ocellaris said:
+1 internet point for Factum.

Some people would try to have others believe that DirectX 12 wasn't even going to exist if it hadn't been for AMD showing everyone the way on like three games that supported it for a small number of users.
Click to expand...

LOL
 
Proprietary APIs? Like the S3 Virge and 3DFX Voodoo and pretty much everything before NV1? :confused:

This ride is broken.
 
Pieter3dnow said:
Intel is part of the Khronos group
Click to expand...

http://techreport.com/news/26682/intel-asked-amd-for-mantle-api-spec

PC World reports that Intel expressed interest in experimenting with AMD's Mantle API. Newly-minted AMD Gaming Scientist Richard Huddy made the revelation during an interview last Friday, noting that Intel had "approached [AMD] for access to the Mantle interface, et cetera."

Intel's request was rebuffed, though. Mantle is still in a closed beta, and AMD doesn't plan on opening the API until version 1.0 is complete
Click to expand...

Intel was refused access to Mantle. The API never made it out of beta and so was never open. Another failed project is all it is.

It would not surprise me if Mantle was based on DirectX and or OpenGL.
 
PRIME1 said:
It would not surprise me if Mantle was based on DirectX and or OpenGL.
Click to expand...
Probably not directly.

The improvements AMD made to increase throughput are a lot like console "native" APIs: programs don't build or stage buffers or other generalized things to be fed through another abstraction layer (you still have to create most of these objects, but they can run far more parallel than DX). It puts more of the responsibility on the developer to manage data and memory resources. In some ways that more resembles older versions of OpenGL than D3D, but it's also different due to being done at a lower level.

It's hard to say who is inspired by what. Console native APIs are older than D3D and so is OpenGL. DX12 has likely been in the works for at least as long as Mantle, and the differences between the two are more likely to be that DX12 more closely resembles Xbox One's DX in certain functionality. Could Mantle have been inspired by Sony's APIs or while working with MS on Xbox One? Of course, and probably was inspired by both.
 
Last edited:
PRIME1 said:
http://techreport.com/news/26682/intel-asked-amd-for-mantle-api-spec



Intel was refused access to Mantle. The API never made it out of beta and so was never open. Another failed project is all it is.

It would not surprise me if Mantle was based on DirectX and or OpenGL.
Click to expand...

???

All the API documentation is open at this point, and you could see all the API functions from literally day 1. You could just dump the .dll's exports and get a complete list.

http://www.amd.com/Documents/Mantle-Programming-Guide-and-API-Reference.pdf


With the documentation, it's actually possible to build the headers yourself and put together a working program. There's no real surprises or mysteries about the API any more. You're (as in likely, literally, you) going to be using it's next incarnation at some point in the future so I don't see how that's failed.


(It unsurprisingly looks pretty close to DX12, with the exception of being a C API vs the C++ COM style of DX12. There's _nothing_ strictly hardware specific in the core API)
 
Last edited:
Factum said:
Again, explain that Skylake and Maxwell is DX12_1 Feature Level...reality dosn't agree with your claims.
Click to expand...

Maybe that means that dx12 isn't that much different than dx11? Since, you know, a lot of older hardware supports it?
 
im still laughing about razor1 having me on ignore.

some of the comments from people in here make we wonder about the future of gaming in general...

A lot of ignorance...
 
I just wonder why my ignore list doesn't work for you know who LOL
 
doesnt work for me either on the same person, for the same reason razor1 mentioned...quoting...

although i do tend to peek, especially when it comes to you know who, his responses are sometimes quite amusing...
 
Is the claim now that Mantle = Vulcan?

Because, then once again, I'd suggest you look at the driver level...(there is a reason Vulcan isn't out yet)

But why ignore the rest of my post?

It's a bit like debatting with superstious people...they will resort to "god of the gaps" in absence of documentation supporting their superstion...
 
Remon said:
Maybe that means that dx12 isn't that much different than dx11? Since, you know, a lot of older hardware supports it?
Click to expand...

I think you need to read a bit up yourself sorry.

DX12 != Direct 12_1 Feature Level, sorry.

DX12FeatureLEvels.png


(Remember to add to that picture that Skylake is 12_1)
 
PRIME1 said:
http://techreport.com/news/26682/intel-asked-amd-for-mantle-api-spec



Intel was refused access to Mantle. The API never made it out of beta and so was never open. Another failed project is all it is.

It would not surprise me if Mantle was based on DirectX and or OpenGL.
Click to expand...

Mantle already adopted how shaders are programmed close to how it is done in DX.

Intel is part of the Khronos group

https://www.youtube.com/watch?v=qKbtrVEhaw8

They got Mantle in an early stage acknowledged in this video.
Intel had access as did Nvidia the youtube video proves this.

You prefer to post stuff which has been refuted for a long time and like the rest you are making assumptions from websites which are wrong or based on false reports.
 
Factum said:
Is the claim now that Mantle = Vulcan?

Because, then once again, I'd suggest you look at the driver level...(there is a reason Vulcan isn't out yet)

But why ignore the rest of my post?

It's a bit like debatting with superstious people...they will resort to "god of the gaps" in absence of documentation supporting their superstion...
Click to expand...

https://www.youtube.com/watch?v=qKbtrVEhaw8

Since it is to hard to watch a video and comprehend what is said how the hell can you even participate in this discussion, all of the FUD you are spreading and keep pretending you know what is really going on are disproved by the official video of the Khronos group and one of BSN. Then you start spreading this BS all over again.

Stop blatantly posting about stuff you don't understand then hope by yelling FUD PR FUD or saying stupid things as you never dealt with an API in your life, would somehow makes sense .

You can not refute facts posted in this video.
 
Last edited:
socK said:
???

All the API documentation is open at this point, and you could see all the API functions from literally day 1. You could just dump the .dll's exports and get a complete list.

http://www.amd.com/Documents/Mantle-Programming-Guide-and-API-Reference.pdf


With the documentation, it's actually possible to build the headers yourself and put together a working program. There's no real surprises or mysteries about the API any more. You're (as in likely, literally, you) going to be using it's next incarnation at some point in the future so I don't see how that's failed.


(It unsurprisingly looks pretty close to DX12, with the exception of being a C API vs the C++ COM style of DX12. There's _nothing_ strictly hardware specific in the core API)
Click to expand...


The entire mantle shader portion and many of the API calls are based on Dx11.x from the xbox one, and parts of it is similar to GNMlib, Mantle isn't new, it takes whats out there on the consoles and made it better.

API's are always evolutionary, so to think one is based off another is true, but Dx12 is more based off of DX11.X and also Mantle, this is why MS was able to create it so fast *in our eyes*

I see it as a fail from a company and financial perspective, because the fruits of spending millions of dollars, gave them very little in the short term and in the long term as well. Are the mantle drivers and experience with creating them going to give AMD a leg up vs the competition with Dx12? Unlikely. Will they have a leg up with Vulkan, unlikely. Its nice that it pushed the industry forward quicker yes, but AMD can't be looking at things like this altruistically, there is always money at the bottom line. What was THEIR benefit of Mantle and only THEIR own.

From an industry perspective its a success, but again, AMD isn't in a position to take on world hunger ;)

About tier levels, AMD is only Tier II but the main difference between Tier II and Tier III is the amount of tiled resources, in the life span of Dx 12 cards we probably won't see the need for the extended tiled resources of Tier III.
 
Last edited:
Pieter3dnow said:
https://www.youtube.com/watch?v=qKbtrVEhaw8

Since it is to hard to watch a video and comprehend what is said how the hell can you even participate in this discussion, all of the FUD you are spreading and keep pretending you know what is really going on are disproved by the official video of the Khronos group and one of BSN. Then you start spreading this BS all over again.

Stop blatantly posting about stuff you don't understand then hope by yelling FUD PR FUD or saying stupid things as you never dealt with an API in your life, would somehow makes sense .

You can not refute facts posted in this video.
Click to expand...


Dude you have no idea of what you are talking about. Checks the dates of when things happened. Write out a time line..... You might actually be able to make a better understanding to relationships with why AMD gave Mantle to the Vulkan committee, Hint there was no where else to go......

Its been 6 months so far and Vulkan still isn't available to the general public since the initial specs of Mantle were given to Vulkan committee, and it looks like Vulkan will come out end of this year, so almost 9 months to modify an existing API, it took MS around the same time to create an API and demo it, with drivers with a full game, all of this was done before Mantle was finalized for release.

What facts are in those videos, you were trying to give timelines with that, AMD's refusal to Intel was done months before that video actually close to a year. Releasing an API after completion and not including industry leaders to help with its creation is not being open, its quite the opposite.

Since you want to get technical, I want you to tell me what this is

Code: 
GR_APPLICATION_INFO appINfo = {};
 appInfo.pAppName = "oneApp";
 appInfo.pEngineName = "oneEngine";
 appInfo.apiVersion = GR_API_VERSION;
 GR_UINT32 gpuCount;
 GR_PYSICAL_GPU gups[GR_MAX_PHYSICAL_GPUS];
 GR_RESULT result = grInitAndEnumerateGpus(&appInfo, nullptr, &gpuCount, &gpus[0]);
 if(result == GR_SUCCESS)
 {FindSutableGPu(gpus, gpuCount);}
Here is a snippit of code
tell me what it is and what its for.

Just a hint this is something very important to Vulkan but can be used in Ogl right now and was available 2 years ago, and yes this is lower level programming even though Ogl isn't a low level API ;)
 
Last edited:
Pieter3dnow said:
You prefer to post stuff which has been refuted for a long time and like the rest you are making assumptions from websites which are wrong or based on false reports.
Click to expand...

Just because you don't like the facts that are posted does not make them wrong or false.

DirectX 12 was in development for years and is now available. Mantle is dead. We can all move on now.
 
razor1 said:
Dude you have no idea of what you are talking about. Checks the dates of when things happened. Write out a time line..... You might actually be able to make a better understanding to relationships with why AMD gave Mantle to the Vulkan committee, Hint there was no where else to go......

Its been 6 months so far and Vulkan still isn't available to the general public since the initial specs of Mantle were given to Vulkan committee, and it looks like Vulkan will come out end of this year, so almost 9 months to modify an existing API, it took MS around the same time to create an API and demo it, with drivers with a full game, all of this was done before Mantle was finalized for release.

What facts are in those videos, you were trying to give timelines with that, AMD's refusal to Intel was done months before that video actually close to a year. Releasing an API after completion and not including industry leaders to help with its creation is not being open, its quite the opposite.

Since you want to get technical, I want you to tell me what this is

Code: 
GR_APPLICATION_INFO appINfo = {};
 appInfo.pAppName = "oneApp";
 appInfo.pEngineName = "oneEngine";
 appInfo.apiVersion = GR_API_VERSION;
 GR_UINT32 gpuCount;
 GR_PYSICAL_GPU gups[GR_MAX_PHYSICAL_GPUS];
 GR_RESULT result = grInitAndEnumerateGpus(&appInfo, nullptr, &gpuCount, &gpus[0]);
 if(result == GR_SUCCESS)
 {FindSutableGPu(gpus, gpuCount);}
Here is a snippit of code
tell me what it is and what its for.

Just a hint this is something very important to Vulkan but can be used in Ogl right now and was available 2 years ago, and yes this is lower level programming even though Ogl isn't a low level API ;)
Click to expand...

Just saying there's a typo in your code :p
 
razor1 said:
Dude you have no idea of what you are talking about. Checks the dates of when things happened. Write out a time line..... You might actually be able to make a better understanding to relationships with why AMD gave Mantle to the Vulkan committee, Hint there was no where else to go......

Its been 6 months so far and Vulkan still isn't available to the general public since the initial specs of Mantle were given to Vulkan committee, and it looks like Vulkan will come out end of this year, so almost 9 months to modify an existing API, it took MS around the same time to create an API and demo it, with drivers with a full game, all of this was done before Mantle was finalized for release.

What facts are in those videos, you were trying to give timelines with that, AMD's refusal to Intel was done months before that video actually close to a year. Releasing an API after completion and not including industry leaders to help with its creation is not being open, its quite the opposite.

Since you want to get technical, I want you to tell me what this is

Code: 
GR_APPLICATION_INFO appINfo = {};
 appInfo.pAppName = "oneApp";
 appInfo.pEngineName = "oneEngine";
 appInfo.apiVersion = GR_API_VERSION;
 GR_UINT32 gpuCount;
 GR_PYSICAL_GPU gups[GR_MAX_PHYSICAL_GPUS];
 GR_RESULT result = grInitAndEnumerateGpus(&appInfo, nullptr, &gpuCount, &gpus[0]);
 if(result == GR_SUCCESS)
 {FindSutableGPu(gpus, gpuCount);}
Here is a snippit of code
tell me what it is and what its for.

Just a hint this is something very important to Vulkan but can be used in Ogl right now and was available 2 years ago, and yes this is lower level programming even though Ogl isn't a low level API ;)
Click to expand...

I don't get what you're getting at with that snippet because that's one the most trivial bits of code possible. It's literally one of the first Mantle functions you'll be calling in a Mantle program. It's pretty close to how you do it in DX - the analog over there being IDXGIFactory::EnumAdapters() before you ultimately create the device.

Not low level or interesting and pretty irrelevant to GL since creating a context is pretty awkward by comparison and you have to jump through some hoops on Windows (like creating an initial dummy context just to grab the function to create the real one later).
 
hmm its for blindness textures;)

and I will leave it at that for you guys to ponder over.




And this is just a snippet, the rest of it you still have to pass through many of the functions just like you will with a low level API. Its possible to do many thing is Ogl, that are similar to the new low level API's, the reason why OGL is being superseded by Vulkan is because of the mess its in right now, not that it isn't capable of doing with Vulkan can do.
 
Last edited:
razor1 said:
hmm its for blindness textures;)

and I will leave it at that for you guys to ponder over.

and this is GL code not Mantle so if you want to say Mantle took for Ogl well here it is.
Click to expand...

What is the point of just making stuff up?

Search 'grInitAndEnumerateGpus' in the Mantle documentation:
http://www.amd.com/Documents/Mantle-Programming-Guide-and-API-Reference.pdf

Whoa imagine that it matches.


mantle said:
grInitAndEnumerateGpus

Initializes the Mantle runtime and enumerates the handles of all Mantle-capable physical GPUs present in the system. Each GPU is reported separately for multi-GPU boards. This function is also used to re-enumerate GPUs after receiving a GR_ERROR_DEVICE_LOST error code.
Click to expand...
 
razor1 said:
hmm its for blindness textures;)

and I will leave it at that for you guys to ponder over.




And this is just a snippet, the rest of it you still have to pass through many of the functions just like you will with a low level API. Its possible to do many thing is Ogl, that are similar to the new low level API's, the reason why OGL is being superseded by Vulkan is because of the mess its in right now, not that it isn't capable of doing with Vulkan can do.
Click to expand...

What are you talking about?

That code is taken from here:

http://stackoverflow.com/a/22858445/1028834

But you failed to even read that comment because you are saying its for bindless texture which was him talking about another link: https://www.khronos.org/assets/uploads/developers/library/2014-gdc/Khronos-OpenGL-GDC-Mar14.pdf
 
razor1 said:
Dude you have no idea of what you are talking about. Checks the dates of when things happened. Write out a time line..... You might actually be able to make a better understanding to relationships with why AMD gave Mantle to the Vulkan committee, Hint there was no where else to go......

Its been 6 months so far and Vulkan still isn't available to the general public since the initial specs of Mantle were given to Vulkan committee, and it looks like Vulkan will come out end of this year, so almost 9 months to modify an existing API, it took MS around the same time to create an API and demo it, with drivers with a full game, all of this was done before Mantle was finalized for release.

What facts are in those videos, you were trying to give timelines with that, AMD's refusal to Intel was done months before that video actually close to a year. Releasing an API after completion and not including industry leaders to help with its creation is not being open, its quite the opposite.

Since you want to get technical, I want you to tell me what this is

Code: 
GR_APPLICATION_INFO appINfo = {};
 appInfo.pAppName = "oneApp";
 appInfo.pEngineName = "oneEngine";
 appInfo.apiVersion = GR_API_VERSION;
 GR_UINT32 gpuCount;
 GR_PYSICAL_GPU gups[GR_MAX_PHYSICAL_GPUS];
 GR_RESULT result = grInitAndEnumerateGpus(&appInfo, nullptr, &gpuCount, &gpus[0]);
 if(result == GR_SUCCESS)
 {FindSutableGPu(gpus, gpuCount);}
Here is a snippit of code
tell me what it is and what its for.

Just a hint this is something very important to Vulkan but can be used in Ogl right now and was available 2 years ago, and yes this is lower level programming even though Ogl isn't a low level API ;)
Click to expand...

I find your comments rather stupid to start with it pretends that AMD was looking for a get rich quick scheme. This and your other comment about them having no place to go but beg MS and Khronos group to please use their "failed" API does not make sense either.

Where as the merits are Mantle did push things and in no way was it meant for AMD hardware only, this is not how AMD operates.

About Vulkan taking so long is no surprise it is ruled over by a committee which means that a lot of peope have to sign of on it. Open-GL never been know for it fast and swift updates it lingers on and hardly make a dent.

That Intel does not get to look at an Alpha version of an API is not that shocking and Intel are certainly not leading in graphics. I do not have a time line of when AMD offered to Khronos group but are you trying to say that Intel might have contributed a thing or two?

I think your code is so nice I would have preferred printf hello world all nicely done with cool indentation and pretty brackets :)

Mantle was done for the purpose of helping game developers that it did for all the reasons that AMD never brought to light about why just a hand full of developers and no other people interfering was prolly done for a good reason.
 
Pieter3dnow said:
Open-GL never been know for it fast and swift updates
Click to expand...

It's updated every year. How often is DX updated?

And unlike DX a vendor can provide an extension for special functionality the very day they release their video card.

Ohh and the user doesn't have to wait around for a special OS to be released just to get it.
 
Those however have been in the hopper for years before they make it to the update.
 
jwcalla said:
It's updated every year. How often is DX updated?

And unlike DX a vendor can provide an extension for special functionality the very day they release their video card.

Ohh and the user doesn't have to wait around for a special OS to be released just to get it.
Click to expand...

If it's so great, why does almost no Windows game use it? All the devs are just stupid, is that it?
 
its really just the defacto standard. Most of the schools teach HLSL as standard for a graphics programming course, microsoft has a lot of documentation readily available and easy-to-obtain, as well as updated.

OGL may be on-par technically, but the money talks when it comes to support and adoption.
 
The money and "just because" arguments always work. MS has money, so it must be evil and doing evil, and they're always "lucky". Alternatively, there is this history: http://programmers.stackexchange.co...do-game-developers-prefer-windows/88055#88055 - which details massive open gl ineptness, internal failures and bad decisions. And also explains that open gl extensions weren't all they were cracked up to be because you had to have AMD and Nvidia specific paths, not just to run at peak efficiency, but to run at _all_.
 
Last edited:
The Mac said:
Those however have been in the hopper for years before they make it to the update.
Click to expand...

Nope.

devil22 said:
If it's so great, why does almost no Windows game use it? All the devs are just stupid, is that it?
Click to expand...

Because there's only one vendor that doesn't have trainwreck OGL drivers. That's a fact.

But I didn't say it was so great... I just countered his idea that it's slow to adopt new features, etc. OpenGL is way ahead of, e.g., DX11. Gamers were throwing away their Kepler cards and never even used its most significant feature (bindless graphics). I guess devs weren't interested in something like that. Then again many are still making DX9 games so oh well.

devil22 said:
The money and "just because" arguments always work. MS has money, so it must be evil and doing evil, and they're always "lucky". Alternatively, there is this history: http://programmers.stackexchange.co...do-game-developers-prefer-windows/88055#88055 - which details massive open gl ineptness, internal failures and bad decisions. And also explains that open gl extensions weren't all they were cracked up to be because you had to have AMD and Nvidia specific paths, not just to run at peak efficiency, but to run at _all_.
Click to expand...

That was ten years ago. The intention of creating DirectX is the same as that of Metal: platform lock-in. If it weren't so, they'd be offered as cross-platform open standards.
 
jwcalla said:
Nope.



Because there's only one vendor that doesn't have trainwreck OGL drivers. That's a fact.

But I didn't say it was so great... I just countered his idea that it's slow to adopt new features, etc. OpenGL is way ahead of, e.g., DX11. Gamers were throwing away their Kepler cards and never even used its most significant feature (bindless graphics). I guess devs weren't interested in something like that. Then again many are still making DX9 games so oh well.



That was ten years ago. The intention of creating DirectX is the same as that of Metal: platform lock-in. If it weren't so, they'd be offered as cross-platform open standards.
Click to expand...

Depends on your point of view. It strikes me as not very different than saying other people are just sucking your air. People like you are just overly cynical. MS might want to make DX12, to make sure their users don't end up with something like open gl of old. Who's to say open gl would even be any better today without MS doing directx in a way that was obviously better since open gl is on windows and developers reject it (which causes it not to get developed, and so on in a cycle.) Anyways doesn't it get tiring trying to second guess MS' motives all the damned time?
 
Pieter3dnow said:
I find your comments rather stupid to start with it *snip*.
Click to expand...

Instead of going all emotional and mental...care to answer some of all the questions you have choosen to ignore?
Trying to find a single argument where you MIGHT appear to have a point...makes no "victory" when you've ignored tenfold more questions...
 
PRIME1 said:
This is why Gameworks is so successful. Game developers won't go the extra mile to enable features like this. However if it's provided to them in a nice usable package they will use it.
Click to expand...

For extra funsies, SLI are regulary broken in gameworks titles... :(

Batman, FC4 and the list goes on......
 
devil22 said:
Depends on your point of view. It strikes me as not very different than saying other people are just sucking your air. People like you are just overly cynical. MS might want to make DX12, to make sure their users don't end up with something like open gl of old. Who's to say open gl would even be any better today without MS doing directx in a way that was obviously better since open gl is on windows and developers reject it (which causes it not to get developed, and so on in a cycle.) Anyways doesn't it get tiring trying to second guess MS' motives all the damned time?
Click to expand...

It's not about second-guessing MS motives. When companies achieve an area of market dominance it is only natural for them to pursue ways to reduce the risk of competition. It's like Intel no longer licensing the x86 ISA. That's a pivotal defense strategy. Likewise, what other operating systems can you use to play DirectX games? Ohh yeah... none. You're stuck with Windows. Seems like that was a pretty good strategy.

The whole argument is off-point anyway. OpenGL is not something developed by some cavemen in some anonymous country somewhere. It's made up of nvidia, AMD, Intel. Apple, Qualcomm, Samsung, etc. are all members. If MS wanted they could also join and offer their additions as well. Hell, the Khronos president is an nvidia employee. It's not like the joint is run by drooling idiots. MS has always been insular and closed off from the community. And Apple is moving in that same direction too now that they've become successful.
 
You must log in or register to reply here.
Back
Top