Hermits, did you guys know this all along? (Nvidia J'accuse!)

Loving the AMD shills in here, can always rely on them to conjure up some factoid about Nvidia "lying" or whatever.

2. To suggest fame pacing is better on AMD than Nvidia is false since its a game to game and driver to driver basis, Now with scaling with XDMA that is a different story, especially with more than two gpus.

4. To suggest AMD is more open source and consumer friendly is a half truth. Only time AMD is willing to open up with their proprietary items like with TressFX or Freesync is when they are hurting or have no choice.Given the choice they would do same things as Nvidia.

. Mantle is a more recent example, AMD saying that Mantle was open to anyone to use but didnt disclose the fact that anyone using their API would have to share all work and data then having to be approved by AMD. That is not open source.

Also Mantle is proprietary since it only used by AMD and their GCN architecture. They promoted Mantle the same as Nvidia promotes Physx or Gameworks etc. Since Mantle wasn't gaining the ground they wanted it to ie increasing their gpu sales and adoption rate and the fact that MS announced DX12, AMD scraped the API since it would die like Glide did back in the day. They decided to give Kronos group a head start on their Modern OpenGL API by giving them most of Mantle's base.

For CES 2016, showing two 980M as Pascal is lying and that's not rocket science. I love NVIDIA shills defending their dear leader.

Fact is you're ignoring AMD gave Mantle to Kronos hence Vulkan was created AND AMD didn't made any promise on Mantle's transfer date which is concluded with AMD transferred Mantle to open standards body i.e. Kronos.

2. Stop being a hypocrite. Should one include NVIDIA's old "bump gate" incidents with their Maxwellv2 product offerings? AMD's XDMA enabled GPUs supercedes their non-XDMA offerings.

4. The only half truth is your post.

Fact is you're ignoring AMD gave Mantle to Kronos hence Vulkan was created AND AMD didn't made any promise on Mantle's transfer date which is concluded with AMD transferred Mantle to open standards body i.e. Kronos.

Mantle's Async compute features wasn't completed for Battlefield 4 PC edition, hence there's very little point giving uncompleted Mantle API to Kronos.

Fact is AMD ASIC GPU documentation is available to 3rd party driver developers for small OS vendors and one can't say the same for NVIDIA GPUs.

lol your totally bypassing and ignoring the point.

Mantle was proprietary since it only used by AMD and their GCN architecture. They promoted Mantle the same as Nvidia promotes Physx or Gameworks etc.

The fact that mantle would have gone out like Glide did back in the day and that AMD gave mantle base to Kronos not to totally throw out years of work has no bearing with the fact that when mantle was mantle under AMD it was proprietary. Which was the whole point of the post ie Mantle was proprietary

Your points are flawed. Fact is you're ignoring AMD gave Mantle to Kronos hence Vulkan was created AND AMD didn't any promise on Mantle's transfer date.

As of this present time/As of today, AMD has transferred Mantle to open standards body, hence Kronos Vulkan was created.

Mantle's Async compute feature wouldn't work on Maxwellv2 i.e. it only has a single command processor with 31 compute queues + 1 graphics queue.

Computer science 101 will tell you a single processor must have SYNC'ed (or mask out the memory location) any multiple data write streams before data read operations i.e. you don't want command processor fetching data that is currently being written by the CPU.

At different memory locations, having multiple command processors enables multiple data reads independent from one another, hence it's Async.

However, NVIDIA's support for their Kelper GOU card dwindled after new series i.e. it's already in legacy support status.

Against NVIDIA Maxwell v1, AMD doesn't need to update their lower GCN 1.0 SKUs.

Yeah, they had frame pacing issues about equal to Nvidia's both technologies (SLI and crossfire) have it still, just crossfire is better now and they were about equal from both technologies inception upto the 7000 series which was just far out. The 7000 series was the anomaly with pretty crappy micro stutter, still, both SLI and Crossfire have micro stutter, but have a look at the difference between the 295x2 micro stutter and Nvidia's in that link i sent, Nvidia's is huge compared to AMD's. It seems pretty obvious to me that crossfire has been superior for years (read: leg up) and that AMD have the better cards in my opinion.

Actually from where i'm standing the single most powerful card out there is still the 295x2. So yeah, swings and roundabouts man, AMD then Nvidia then AMD then Nvidia. All the way back to the days of All in wonder and TNT cards. It just depends on when you look at things.

As for rebrands, so what, Nvidia has been rehashing cards as well, do you think they don't do the same thing or something?? GeForce 9800 -> GTS 250, GeForce 405 card is a rebranded GeForce 310 which itself is a rebranded GeForce 210. In fact:

• The GeForce 605 (OEM) card is a rebranded GeForce 510.
• The GeForce GT 610 card is a rebranded GeForce GT 520.
• The GeForce GT 620 (OEM) card is a rebranded GeForce GT 520.
• The GeForce GT 620 card is a rebranded GeForce GT 530.
• The GeForce GT 630 (DDR3, 128-bit, retail) card is a rebranded GeForce GT 430 (DDR3, 128-bit).
• The GeForce GT 630 (GDDR5) card is a rebranded GeForce GT 440 (GDDR5).
• The GeForce GT 640 (OEM) card is a rebranded GeForce GT 545 (DDR3).
• The GeForce GT 645 (OEM) card is a rebranded GeForce GTX 560 SE.
• The GeForce GT 705 (OEM) is a rebranded GeForce GT 620 (OEM), which itself is a rebranded GeForce GT 520.
• The GeForce GT 730 (DDR3, 64-bit) is a rebranded GeForce GT 630 (Rev. 2).
• The GeForce GT 730 (DDR3, 128-bit) is a rebranded GeForce GT 430 (128-bit).
• The GeForce GTX 760 Ti (OEM) is a rebranded GeForce GTX 670.

This isn't exactly a new thing for Nvidia, they have been rebranding cards for the last half a decade.

http://www.alphr.com/blogs/2010/02/24/its-at-it-again-nvidia-rebrands-yet-more-gpus

AMD does it too, who cares as long as there are improvements what does it matter to anyone.

I heard they gave Steve Jobs cancer and secretly poisoned walt disney :O

Those @ssholes!!!

Proprietary means they owned it or operated it under a registered trade name, not quite the same as saying that they restricted it to only work with the GCN architecture, that's not what proprietary means. AMD designed something that anyone could play with that used there own GPU architecture that's not that uncommon a thing for people to do you know and hardly a bad thing.

But it was indeed open source. Nvidia is most definitely not open source. Not that this really matters too much.

Open source can still mean that AMD have the last say on what updates get distributed through there Mantle, what it also means is that anyone who wanted to make there own update and publish it somewhere can too.

Actually he's quite wrong here, AMD have supported the 6000 series right up to the 3rd quarter of 2015. I was still using my 6970's at that point. And still supporting the 5990 as far as i'm aware until just recently (friend still uses one).

@jereb31: No it's stupid to include multi GPU's but if it makes you feel better about AMD, go right ahead. I have no dog in this fight.

On paper and in benchmarks, AMD has had the edge on pure performance in terms of having the fastest single card a few times but in games, the dual GPU cards lose that edge.

@Juub1990:

Doesnt seem stupid to me. If someone is making the statement that Nvidia was unrivalled in the gpu sector for the last 3 years deapite not actually having the most powerful gpu for even 2 in a row.

Seems to me your dog is Nvidia.

But its just more information. Do with it what you want im obviously not going to convince you.

No they had the most powerful GPU for about 30 months out of the last 36.

You can include multi-GPU all you want. I choose not to because it's a completely different beast.

I have no dog in this fight. These are just facts. If you include the 7990 then we have to count the 690 which was pretty much equal until the wonder drivers then you have to include SLI 780 Ti, Titan Z etc.

@ronvalencia:

Wrong its no longer Mantle aka under amd's control , hence you still missing the point

This is because nvidia and amd have to major differences in marketing strategy. AMD focuses in open source to implement new features as an open-standard in hopes that users use the standard readily available technology. They also target to make PC gaming as a whole better (Vulkan/as that would drive more PC gamers in general. Nvidia focuses on making their technology exclusives so that current PC gamers would pick them over any rivals. I don't see anything wrong with the premise of what NVIDIA is doing (it is their execution some times that is horrendous).

Wrong on all accounts, Mantle was not "open" nor was it Open source. Open source means "denoting software for which the original source code is made freely available and may be redistributed and modified." Fact is that when Mantle was under AMD control it was Proprietary ie only AMD used it and only worked on their products , even if AMD allowed Intel and Nvidia to use it, AMD would have the final word in how it was used and required the others to disclose all plans and usage. Unless they changed it and it became open source, in which Nvidia and Intel could used the base of Mantle, modified it and claimed their own versions of the API just like Vulkan.

AMD updates and fixes for their dual gpu cards dwindled with dual 4000 series ( i know someone that had a 4870X2) and dual 5000 series have suffered and even the 6000 series less to a degree but still suffered. fact that they totally stopped supporting the 5000 and 6000 series even though they are DX11 based gpus is bad.

Just had more of a dig and your right about not being open source, but it was in fact free ware. Still allowing everyone else to use it is quite a lot better than not allowing anyone else to use it wouldn't you say so?

It didn't dwindle at all, i also had the 4870x2, used it up until 2011 when i bought the 6970's. Worked fine (ran hot) but was supported right up until 2 generation later and probably then some. How can you quantify that the support has been dwindling when they are still releasing driver updates for these cards generations after there release?

So the HD4000 series of graphics cards came out in 2008, they received updates until Catalyst 13.1 2013. That's a good 5 years, i wouldn't call that dwindling support.

It's not a different beast at all, fact is AMD have the most powerful card on the market for less than half the price of Nvidia's. Doesn't matter if it's a dual GPU or not. You can totally include all other SLI and crossfire configs, fact is still that AMD has the most powerfull card for less it doesn't change that.

http://www.extremetech.com/gaming/220818-nvidias-drive-px-2-prototype-allegedly-powered-by-maxwell-not-pascal

These issues with Pascal and the Drive PX 2 echo the Fermi “wood screw” even of 2009. Back then, Jen-Hsun held up a Fermi board that was nothing but a mock-up, proclaimed the chip was in full production, and would launch before the end of the year. In reality, NV was having major problems with GF100 and the GPU only launched in late March, 2010.

....

The entire point of holding up a product from stage is to demonstrate to the audience that the hardware actually exists. If it’s later shown that the hardware in question wasn’t what it was claimed to be, it undercuts the original point. Worse, it invites the audience to question why the company is playing fast and loose with the truth. There’s no reason to think Pascal is suffering an unusual delay — but these kind of antics invite speculation to the contrary.

Wrong. Fact is you're ignoring AMD gave Mantle to Kronos hence Vulkan was created AND AMD didn't any promise on Mantle's transfer date, hence your point is flawed.

As of this present time/As of today, AMD has transferred Mantle to open standards body, hence Kronos Vulkan was created.

If Intel or NVIDIA are interested with Mantle, they can create their own Mantle API to DirectX12 API bridge.

https://www.gamedev.net/topic/666419-what-are-your-opinions-on-dx12vulkanmantle/#entry5215019

The Mantle spec is effectively written by Johan Andersson at DICE, and the Khronos Vulkan spec basically pulls Aras P at Unity, Niklas S at Epic, and a couple guys at Valve into the fold.

Three out of those four just made their engines public and free with minimal backend financial obligation.

Your AMD control over Mantle argument is a red herring. For Mantle spec, blame EA-DICE. For Vulkan spec, blame EA-DICE, Unity, Epic and Valve. EA's involvement into computer platform's behaviour relates back to Commodore-Amiga (1985) and 3DO (1993). Valve's Steam machine is effectively 3DO version 2.0 i.e. game console standard that can be cloned.

AMD just follows it's loudest users i.e. mainstream 3D engine developers.

Well, that pretty much settles it, I will definitely not be putting any nVidia hardware in my next computer.

Been a Radeon guy all along, and will remain.

Wrong... not ignoring that Kronos has Mantle base turned into Vulkan your ignoring what was posted about what Mantle was.... its proprietary nature during its usage before it was abandoned by AMD.

Freeware is different than open source, and free ware again cant be modified, re-distributed or re-engineered without author's permission. There is a reason why Intel and Nvidia didnt bother even asking to use Mantle because AMD again has the final word in how its used.

you are mixing up general driver support with the drivers vs not seeing the proper fixes and enhancements for those dual gpu's. Every new gen of a gpu released those older dual gpus suffered more so than the singles.

Mantle was always an open standard from the beginning. AMD's implementation for GCN is proprietary closed source. People that claim Mantle is open source don't know what mantle is. NVidia/Intel were always free to use the API they just have to write specific driver implementations that use their specific hardware.

Wrong. A view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

https://www.gamedev.net/topic/666419-what-are-your-opinions-on-dx12vulkanmantle/#entry5215019

The Mantle spec is effectively written by Johan Andersson at DICE, and the Khronos Vulkan spec basically pulls Aras P at Unity, Niklas S at Epic, and a couple guys at Valve into the fold.

Three out of those four just made their engines public and free with minimal backend financial obligation.

Your AMD control over Mantle argument is a red herring. For Mantle spec, blame EA-DICE. For Vulkan spec, blame EA-DICE, Unity, Epic and Valve.

AMD's binary blob that follows Mantle spec is close source, but the API spec for Mantle is not proprietary since AMD did not create it i.e. it's Johan Andersson at EA-DICE.

@ronvalencia:

Again ignoring the facts at hand Mantle was proprietary, AMD controlled all the aspects in how its used if it was open to other gpu architectures DICE would have used it. dont use DICE as the excuse why Mantle only worked with GCN....AMD promoted it lol

"The key targets for AMD when it comes to Mantle were higher framerates, reduced rendering latency, reduced GPU power consumption, better use of multi-core CPUs, and re-pioneering new features like split-frame rendering. Many of those targets have been achieved. A number of games that use Mantle give advantages to AMD Radeon graphics adapters."

"proprietary API costs money to develop and maintain. If there is an open-standard cross-platform API that can provide AMD the same advantages as Mantle, then the latter is simply not something that money-starving AMD should invest in. Apparently, AMD worked closely with Khronos and many parts of Mantle became parts of Vulkan."

"it looks like from now on Mantle will become an API for select game developers that collaborate with AMD and want to use exclusive capabilities of Radeon hardware before they are supported by Vulkan, OpenGL or DirectX. In short, future versions of Mantle will become even more proprietary than the current iteration because they will not be available for all."

The upgrade itch is killing me, and if rumors are true, I might end up going with Polaris if that comes out first, though the reasonable thing would be to wait for Pascal for the comparisons. I do feel that AMD might have an advantage this time, let's see how NVIDIA deals with building a compute oriented architecture and maintaining what Maxwell is known for: efficiency.

I still have my 7970, terrific card, I consider it the 2500K of GPUs. I've even seen benchmarks where it goes neck and neck with the 780 and the 3GB VRAM was just the icing on the cake (RIP GTX 680). Curious how I no longer see that many people with 600 series cards, while there are still quite a few gamers holding on to their 7970, 7870 (this last one I've seen it go neck and neck with the 680 as well).

It's also funny that the Gimpler debacle blew up until The Witcher 3 released when it had been going on for months.

Sucks to be an early adopter, but like I said, I'm set for AMD if all things go well, but I wouldn't like to see a 490X "GHz Edition" a few months later from launch. I'm hoping for an April/May release so I can pickup one around July when all of the suckers have picked up the defective units.

Again, you are ignoring the facts. EA-DICE has stated Mantle API is open for other vendors and AMD hasn't made any promise on the transfer date.

Vulkan's existence is evidence of the Mantle's transfer and multi-vendor goals. Mantle's graphics API solution has to be proven first without multi-vendor politics that crippled OpenGL's development.

Your arguments are against the primary source and the current situation which makes your point flawed.

Furthermore, NVIDIA has it's own path for reducing overheads i.e. a vendor specific API extension kit-bash to legacy OpenGL and you chose to ignore this.

In the Kronos meeting, NVIDIA's OpenGL + vendor specific extensions was out-voted and AMD/EA-DICE's Mantle API solution wins as the new basis for GL/Next hence Vulkan. The majority of Kronos's voting membership doesn't want NVIDIA's OpenGL + vendor specific extensions.

Effectively, EA-DICE, Unity, EPIC and Valve spec'ed Vulkan APIs with improvements on cross vendor compatibility and remove Microsoft Windows dependency.

Intel, ARM**, Imagination Technologies** (for PowerVR), Qualcomm** (for dominant Phone SoC market share) and AMD** would have out-voted NVIDIA's solution.

**Same companies are also in HSA.

Actually 04dcarraher is right. It was "intended" to be used openly but it started out completely in AMD control, died miserably when dx12 was announced and amd started warning its customers to back off of mantle, and vulkan is.. not mantle, just the best parts were taken from mantle when creating vulkan.

At least someone sees what Im talking about

Actually, 04dcarraher is not correct. A 100 percent AMD controlled stack wouldn't be "Mantle" as AMD has other lower level runtime solutions. Mantle's intention to be an open standard is Vulkan's existence.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

That's nearly 1 : 1 remapped.

Mantle's MS HLSL has to be removed since it's Microsoft's IP. PS4 has MS HLSL to PSSL converter tool to get around this problem.

AMD was betting on MS's old "assimilate and extend" tactics to defend MSFT's Windows runtime ecosystem. Direct3D12 was also based from Mantle.

There's another "Mantle" fork and it's called DirectX12.

Note why XBO's DirectX11.X wasn't the basis for DirectX12.

The ultimate goal for the graphics API change was for AMD's CPU single thread issue.

For the PC, AMD could have released DX11 with a working MT model from XBO, but that would not completely solve AMD's CPU issues.

@ronvalencia said:

Actually, 04dcarraher is not correct. A 100 percent AMD controlled stack wouldn't be "Mantle" as AMD has other lower level runtime solutions. Mantle's intention to be an open standard is Vulkan's existence.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

hat's nearly 1 : 1 remapped.

Mantle's MS HLSL has to be removed since it's Microsoft's IP. PS4 has MS HLSL to PSSL converter tool to get around this problem.

AMD was betting on MS's old "assimilate and extend" tactics to defend MSFT's Windows runtime ecosystem. Direct3D12 was also based from Mantle.

There's another "Mantle" fork and it's called DirectX12.

Note why XBO's DirectX11.X wasn't the basis for DirectX12.

The ultimate goal for the graphics API change was for AMD's CPU single thread issue.

For the PC, AMD could have released DX11 with a working MT model from XBO, but that would not completely solve AMD's CPU issues.

Mantle IS proprietary. It might have been intended to be open but NVidia/Intel didn't have any say in forming it's shape. It also was explicitly discontinued...

Vulkan IS NOT mantle. You can't post 20 commands that are similar when it's been verified across the web that Vulkan is only using parts of mantle. If you'd like to go map the rest of the commands from the 450 page developer guide then you might be on to something.

DirectX 12 is a similar technology. It's not a fork of mantle that's just stupidity. Its similar because they both target a similar usages for developers.

Still freeware is a shite side better than closed in this case.

Again how can you quantify dwindling support. I received constant updates when I used it, including bug fixes right up until i upgraded in 2011. That is not dwindling support. On this note has Nvidia been dwindling in there support of multi GPU/SLI configs to the same degree.

It sounds like you are just trying to highlight generational driver creep where the old ones all receive less support than the new ones. Dual GPU's of the same generation still benefit from the same GPU driver updates of that generation (they are still the same GPU's after all), but bundled in a pre-crossfired package.

Fact is "Mantle" could only be used by AMD's GCN architecture which means that it was proprietary. Fact that it was suppose to be open for anyone to use and that Vulkan is using many aspects and parts of mantle has nothing to do with the fact that "Mantle" when it was under AMD's control and was still being promoted and developed as "Mantle" only AMD products could use it. ie meaning proprietary, just like Nvidia's Physx or G-sync

The main difference being that no one had to pay for a license to use it, whilst with Nvidia's G-Sync you did.

@waahahah said:

@ronvalencia said:

Actually, 04dcarraher is not correct. A 100 percent AMD controlled stack wouldn't be "Mantle" as AMD has other lower level runtime solutions. Mantle's intention to be an open standard is Vulkan's existence.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

hat's nearly 1 : 1 remapped.

Mantle's MS HLSL has to be removed since it's Microsoft's IP. PS4 has MS HLSL to PSSL converter tool to get around this problem.

AMD was betting on MS's old "assimilate and extend" tactics to defend MSFT's Windows runtime ecosystem. Direct3D12 was also based from Mantle.

There's another "Mantle" fork and it's called DirectX12.

Note why XBO's DirectX11.X wasn't the basis for DirectX12.

The ultimate goal for the graphics API change was for AMD's CPU single thread issue.

For the PC, AMD could have released DX11 with a working MT model from XBO, but that would not completely solve AMD's CPU issues.

Mantle IS proprietary. It might have been intended to be open but NVidia/Intel didn't have any say in forming it's shape. It also was explicitly discontinued...

Vulkan IS NOT mantle. You can't post 20 commands that are similar when it's been verified across the web that Vulkan is only using parts of mantle. If you'd like to go map the rest of the commands from the 450 page developer guide then you might be on to something.

DirectX 12 is a similar technology. It's not a fork of mantle that's just stupidity. Its similar because they both target a similar usages for developers.

The FACTS.

1. AMD and EA-DICE has the intention to make Mantle available for non-AMD GPU vendors.

2. AMD and EA-DICE made no promises on the transfer date.

3. AMD transferred Mantle API spec to Khronos.

4. Khronos modifies Mantle to improve cross vendor compatibility, hence Vulkan aka Volcano which follows AMD's Earth name themed marketing. "OpenGL 5.0" overridden by AMD's marketing names.

5. NVIDIA has their own solution with their with vendor specific OpenGL extension kit-bash.

The reason for similarity between DirectX12 and Mantle is mostly due to the similar activist players e.g. EA-DICE (effectively the author for Mantle API spec), AMD (for MS XBO) and Mantle's rendering programming model was a proven solution that has multiple 1st tier games.

@waahahah:

You can't post 20 commands that are similar when it's been verified across the web that Vulkan is only using parts of mantle. If you'd like to go map the rest of the commands from the 450 page developer guide then you might be on to something.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

To keep things into perspective

1. NVIDIA's PhysX doesn't have the same open status as Bullet Physics.

2. NVIDIA's Gameworks doesn't have the same open status as AMD's GPUopen.

3. NVIDIA's G-Sync doesn't have the same open standard as AMD's FreeSync.

4. NVIDIA's CUDA+ARM fusion programming model doesn't have same open status as AMD's HSA.

5. AMD/EA-DICE's Mantle API spec was transferred to open standards body Khronos. NVIDIA's OpenGL+vendor specific extension kit-bash alternative future was dumped by the majority of Khronos' voting members. AMD can rely on HSA members to vote on their side.

The facts are also

intention to make open != open, it was a proprietary API

AMD DISCONTINUED mantle.

AMD transfered to Khronos since... it was DOA

Khronos is not using the mantle api, some of it was discarded, more was added. Like any group who aqcuire someone else's technolog, the pick pieces that they think will be useful and discard the rest.

It's no longer AMD's technology, yes something good came about AMD's failure because of DX12 or NVidia's inability to adopt mantle earlier on. But it's still a failure caused by proprietary licensing.

Of course similar activist were involved with directx, that doesn't make it a fork. There really isn't that many big players in the industry, and more then likely dx12/mantle were developed simultaneously to solve the same problems, who would have thunk it came out to similar solutions. You could just as well make the argument that dx12/mantle were both a result of MS/AMD work on DX for xbox one. They may have made it to market first but it because completely redundant technology that likely wouldn't get used with DX12 on the horizon.

also, 12 games support mantle. Not many people touched it.

https://en.wikipedia.org/wiki/Category:Video_games_that_support_Mantle

The FACTs are

1. The intention to open Mantle API was fulfil by transferring Mantle API to Khronos.

2. Mantle API's focus was discontinued after Mantle API was transferred to Khronos. Mantle is alive for advance prototyping and testing.

3. AMD's Mantle binary blob remains close source e.g. AMD's Mantle API/MS HLSL LLVM engine remains close source.

4. Mantle has touched major 3D engine developers and it's programming model lives on with Vulkan.

@waahahah:

Khronos is not using the mantle api, some of it was discarded, more was added. Like any group who aqcuire someone else's technolog, the pick pieces that they think will be useful and discard the rest.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

Who are you again?

Btw, Mantle's MS HLSL must be removed from any open standards, hence Mantle can't exist "as is" in any open standards body.

1. Not a fact. It was a proprietary API intended to be open, transfered to Khronos after dieing. They intended for a public API that they owned. It wasn't intended to be transfered to Khronos until it was DOA.

https://community.amd.com/community/gaming/blog/2015/05/12/on-apis-and-the-future-of-mantle

Mantle’s definition of “open” must widen. It already has, in fact. This vital effort has replaced our intention to release a public Mantle SDK, and you will learn the facts on Thursday, March 5 at GDC 2015.

2. Mantle was discontinued and transfered to Khronos? Does it matter in which order? It died, Khronos got the carcass.

3. Vulcan is a true fork of mantle at this point, and its different. They didn't pick all of it up, they added a lot of new stuff. Its no longer mantle.

The only reason Khronos got it is because it makes their and AMD's life a little easier. Both sides have less work to do. AMD Lost either way because they held all the decision making to the API. Just because it was open for people to use doesn't mean they'll accept changes NVidia/Intel wants. Also, 12 games.

I see the key phrase "was a proprietary API".

1. Notice "It already has" after "Mantle’s definition of “open” must widen".

Mantle's focus was changed after Vulkan was created. Mantle did NOT die before the transfer to Khronos.

Vulkan's creation has fulfil Mantle's open standards intention. Remember, the competitor to Mantle was NVIDIA's vendor specific OpenGL extension kit-bash.

2. For the narrative, the order matters i.e. it changes point 1 and you made an error in point 1.

3. Again, Mantle API can't exist "as is" e.g. MS HLSL must be removed from in any open standards body. Mantle must change to support reduced feature GPUs e.g. NVIDIA Kepler and Maxwellv1, Intel Haswell and Broadwell IGP and 'etc'.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

As an example, UAV slots for all shader stage is a hardware issue and API has to be aware with GPUs that doesn't have full heap features.

@waahahah said:

AMD Lost either way because they held all the decision making to the API.

As stated by ex-NVIDIA driver developer, Mantle API spec was effectively authored by EA-DICE's Johan Andersson.

In 2011, Electronic Arts was the world's third-largest gaming company by revenue after Nintendo and Activision Blizzard.

@ronvalencia said:

1. That's not a fact. Mantle's focus was changed after Vulkan was created. Mantle did NOT die before the transfer to Khronos.

2. For the narrative, the order matters i.e. it changes point 1 and you made an error in point 1.

3. Again, Mantle API can't exist "as is" e.g. MS HLSL must be removed from in any open standards body. Mantle must change to support reduced feature GPUs e.g. NVIDIA Kepler and Maxwellv1, Intel HD 4000 and 'etc'.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

The order does not matter one single bit. The fact is we don't know the true order. We see it the way the PR spun it. The link i gave was AMD discontinuing prior to their announcement of transferring it to Khronos. At the end of the day only a handful of EA games and star citizen were supporting it. The API was DOA outside of EA. It was not a widely used technology and was not going to be while in the hands of AMD. Mantle 1.0 was released... and that was it, the public API known as mantle discontinued and the open API Vuclan began. Someone in amd one day looked at the support they were getting and said "shit this isn't going to work". If they wanted their hard work to pay off, they'd have to give it to a impartial 3rd party.

You can argue about semantics all you want, before or after, the reasoning for discontinuing is still the same. There wasn't enough adoption of mantle to make it viable. AMD gave it up. And now OpenGl has an open API that they are changing to meet the needs of multiple vendors. It's no longer mantle. Mantle ended at 1.0, vulkan will continue to change and evolve as an Open API.

@waahahah said:

@ronvalencia said:

1. That's not a fact. Mantle's focus was changed after Vulkan was created. Mantle did NOT die before the transfer to Khronos.

2. For the narrative, the order matters i.e. it changes point 1 and you made an error in point 1.

3. Again, Mantle API can't exist "as is" e.g. MS HLSL must be removed from in any open standards body. Mantle must change to support reduced feature GPUs e.g. NVIDIA Kepler and Maxwellv1, Intel HD 4000 and 'etc'.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

The order does not matter one single bit. The fact is we don't know the true order. We see it the way the PR spun it. The link i gave was AMD discontinuing prior to their announcement of transferring it to Khronos. At the end of the day only a handful of EA games and star citizen were supporting it. The API was DOA outside of EA. It was not a widely used technology and was not going to be while in the hands of AMD. Mantle 1.0 was released... and that was it, the public API known as mantle discontinued and the open API Vuclan began. Someone in amd one day looked at the support they were getting and said "shit this isn't going to work". If they wanted their hard work to pay off, they'd have to give it to a impartial 3rd party.

You can argue about semantics all you want, before or after, the reasoning for discontinuing is still the same. There wasn't enough adoption of mantle to make it viable. AMD gave it up. And now OpenGl has an open API that they are changing to meet the needs of multiple vendors. It's no longer mantle. Mantle ended at 1.0, vulkan will continue to change and evolve as an Open API.

From https://community.amd.com/community/gaming/blog/2015/05/12/on-apis-and-the-future-of-mantle

MARCH 04, 2015: Did you know that the Khronos Group has selected Mantle to serve as the foundation for Vulkan, a low-overhead PC graphics API that works on multiple OSes and hardware vendors

Your link doesn't support your "AMD discontinuing prior to their announcement of transferring it to Khronos" argument.

From https://en.wikipedia.org/wiki/Vulkan_(API)

The Khronos Group began a project to create a next generation graphics API in July 2014 with a kickoff meeting at Valve Corporation.^[18] At SIGGRAPH 2014 the project was publicly announced with a call for participants.^[7]

According to the US Patent and Trademark Office, the trademark for Vulkan was filed on February 19, 2015

SIGGRAPH 2014 was held on December 3-to-6, 2014

Since Vulkan is part of AMD's Volcano/Earth themed codename and marketing.

NVIDIA uses a different codename marketing e.g. they are based on inventor's names e.g. Kelper, Maxwell, Pascal and 'etc'.

Khronos Group's selection for Mantle have occurred around February 19, 2015.

Timeline Summary

July 2014: Khronos Group began a project to create a next generation graphics API at Valve.

December 3, 2014: At SIGGRAPH 2014 the project was publicly announced with a call for participants.

February 19, 2015: The trademark for Vulkan.

MARCH 04, 2015, AMD's official announcement that Mantle was selected to replace OpenGL.

Between July 2014 to before February 19, 2015, AMD transferred Mantle API spec to Khronos Group.

ROFL, That's not the article, it's an update for the article post vulkan announcement, march 3... and 2 days after the article... march 2. I feel like i won the internet today <3

And like I keep saying, it doesn't matter. I thought you'd get that with me iterating it over multiple times, Obviously AMD new mantle wasn't going anywhere fast. They negotiated with Khronos long before any announcements were made. Any of AMD's partners using the technology would have probably known before hand too, since there were likely getting resources to transfer to vulkan.

From https://en.wikipedia.org/wiki/Vulkan_(API)

The Khronos Group began a project to create a next generation graphics API in July 2014 with a kickoff meeting at Valve Corporation.^[18] At SIGGRAPH 2014 the project was publicly announced with a call for participants.^[7]

According to the US Patent and Trademark Office, the trademark for Vulkan was filed on February 19, 2015

SIGGRAPH 2014 was held on December 3-to-6, 2014

Since Vulkan is part of AMD's Volcano/Earth themed codename and marketing.

NVIDIA uses a different codename marketing e.g. they are based on inventor's names e.g. Kelper, Maxwell, Pascal and 'etc'.

Khronos Group's selection for Mantle have occurred around February 19, 2015.

Timeline Summary

July 2014: Khronos Group began a project to create a next generation graphics API at Valve.

December 3, 2014: At SIGGRAPH 2014 the project was publicly announced with a call for participants.

February 19, 2015: The trademark for Vulkan.

March 4, 2015, AMD's official announcement that Mantle was selected to replace OpenGL.

Between July 2014 to before February 19, 2015, AMD transferred Mantle API spec to Khronos Group.

You lost.

No I didn't, again, timeline doesn't matter. They didn't start transfering it until the nearing the first release. They're partners for the technology never expanded beyond EA, and RSI... they at least got 2 developers on board though!

12 games.

No wide spread support.

Mantle was dead before it was released.

No Intel/NVidia support as AMD proprietary "public" api.

http://techreport.com/news/25651/mantle-to-power-15-frostbite-games-dice-calls-for-multi-vendor-support

Didn't even make it to 15...

No, timeline does matter. Again, February 19, 2015: The trademark for Vulkan.

"Peer review" and selection process must involve Mantle API spec transfer i.e. members has to know what they are voting for.

Where's your source for "They didn't start transfering it until the nearing the first release" assertion?

@waahahah said:

12 games.

So what? That's from https://en.wikipedia.org/wiki/Category:Video_games_that_support_Mantle

It doesn't include 3DMarks and Oxdie's Star Swarm benchmarks.

@waahahah said:

they at least got 2 developers on board though!

They have more than two developers on board.

Sniper Elite 3's devs are not part of RSI nor EA.

I don't recall SQ's Thief 2014 being part of EA or RSI....

@waahahah said:

No wide spread support.

Who cares? NVIDIA's kit-bash vendor specific OpenGL extensions are out and it's related programming model with it.

@waahahah said:

No Intel/NVidia support as AMD proprietary "public" api.

Again, a view from Unity3D's Aras Pranckevičius http://aras-p.info/blog/2015/03/13/thoughts-on-explicit-graphics-apis/

Khronos Vulkan was announced (which is very much AMD Mantle, improved to make it cross-vendor).

.

Im an nvidia fanboy atm, just because i really like the power of using nvidia inspector and other things like dsr and physx and maybe slightly better tessellation performance in the nvidia cards.

But even saying that, Nvidia is very dishonest and hurtful to the gaming market. They don't share their tech so we are now left with physx and gsync stuck on one card. Their gameworks included in games hurts AMD and the gamer directly since it cuts out half the gaming community with this practice. Nvidia goes out of their way to make games run worse on AMD hardware, by either forcing their gameworks into games or making game developers over use tessellation effects, which mid to lower end AMD cards perform worse at.

While AMD is the complete opposite of nvidia, AMD are the good guys, they share their tech, freesync, tressfx, and even Vulkan which AMD has something to do with and isn't hogging it to themselves and its possibly more powerful than DX12 will ever be in theory. Im missing a lot here, AMD has created a lot of tech and shared it while nvidia hogs everything to themselves hurting the PC gaming community. Even AMD and Intel somewhat cooperated on their x64 x86 tech to allow each other to use it.

I really should be an AMD fanboy but theres just a few things in nvidia cards which AMD don't have or do as well. I even fell for their 3.5 GB 970 cards which nvidia listed fraudulently as 4gb cards. AMD has never done anything like this that i know of. Nvidia really does whatever it wants. If AMD wasn't around to keep them in check, we'd all be in a lot of trouble.

And i think i agree with the OP, i don't really want to support nvidia much anymore and may skip out next gen of their cards and wait for something from AMD.