AMD launching open-source GameWorks alternative; GPUOpen

TaintedSquirrel

[H]F Junkie
Joined
Aug 5, 2013
Messages
12,691
http://www.tomshardware.com/news/amd-gpuopen-open-source-development,30750.html

When AMD launches GPUOpen this coming January, it plans to provide access to TressFX 3.0, GeometryFX, AOFX, ShadowFX, a handful of tools, the LiquidVR SDK, DirectX 11 and 12 code samples, compute tools, and several other SDKs. The company is emphatic about its dedication to open source, too. All of the software available through GPUOpen is provided without restriction, per the MIT License.
 
Well, better late than never, I guess.

Hopefully this will help push graphical fidelity in games, while keeping high performance.
 
Those are some pretty red slides.

In all seriousness, I think the Linux portion of this presentation (AMDGPU) is particularly relevant, if Valve wants SteamOS to get any traction.
 
Those are some pretty red slides.

In all seriousness, I think the Linux portion of this presentation (AMDGPU) is particularly relevant, if Valve wants SteamOS to get any traction.

SteamOS doesn't need AMD. Its the other way around.
 
AMD will make these resources available under the permissive MIT License, meaning that GPUOpen resources can be examined, modified, reused, and resold without restriction.

Does that mean Nvidia can use the resources and turn then into their own propietary technology?
 
Awesome, wish they would have shown some examples of all of the effects though.
 
We will see but, I am not going to get excited over something that is not even out yet. Also, open has not worked all that well for AMD based on their financials. I am not an Nvidia fan and would prefer they be more open but, it is not what the industry and gamers want based upon Nvidia's financials. (I have a 980Ti now because otherwise, I would have had to replace my case for the R9 Fury.)
 
Does that mean Nvidia can use the resources and turn then into their own propietary technology?

Here's the MIT license for you

The MIT License (MIT)

Copyright (c) <year> <copyright holders>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
We will see but, I am not going to get excited over something that is not even out yet. Also, open has not worked all that well for AMD based on their financials. I am not an Nvidia fan and would prefer they be more open but, it is not what the industry and gamers want based upon Nvidia's financials. (I have a 980Ti now because otherwise, I would have had to replace my case for the R9 Fury.)

Just means that AMD sponsored titles will run just fine on your new Nidia video card.
 
http://blog.mecheye.net/2015/12/why-im-excited-for-vulkan/


http://www.extremetech.com/gaming/2...ls-an-open-source-answer-to-nvidias-gameworks




A couple takeaways from those blurbs.

One, it will be harder to cheat with vulkan / open source suites for performance like this amd initiative, since the specific tweaks can be modified by any vendor to their own needs.

Having a collection of graphical effects that can be specifically tailored to different hardware as needed is intrinsically more neutral and harder to cheat with compared to something like Gameworks where amd is not allowed to see the black box source code.

Two, I am not at all sure this will stem the tide of lazy/ethically compromised/greedy game developers from sticking with gameworks for the marketing push and other financial reasons. They seem perfectly willing to fuck over stable performance to chase the gameworks branding.


assassins creed unity - got that gameworks, but ran like glitchy ass
batman arkham knight - got that gameworks, but ran like ass, and still has severe constraints
witcher 3 - got that hairworks, runs like ass with it, but not as bad as other titles, just don't you dare have an amd card or an older nvidia card.


The first two game devs would have been better served putting more effort into the core game rather than try to slap on duct tape and sparkle, lipstick on a pig of pc gaming performance, but they don't actually give a shit about the experience, it's about the marketing dollars.


I don't want that to be the case, but most of the people slapping gameworks into their game don't seem to give a crap about the tech actually running better ON TOP of looking better.


It would be better for the entire gaming world if everyone switched to the AMD tools, then no one would be singled out so heavily, and if they were, it would be far more transparent what was going on, instead of nvidia playing hide the ball with shit optimizations and rendering choices done solely to privilege their own latest and greatest.
 
AMD now releases more marketing slides than interesting products.

Won't see a new product until 2016 sometimes. At least they are trying to give you an idea of what they are working on for 2016. Well that's how I perceive it.
 
Won't see a new product until 2016 sometimes. At least they are trying to give you an idea of what they are working on for 2016. Well that's how I perceive it.

AMD's big push for 2016 is going to be getting a bunch of new "open source" features working for Direct X 11? And if they really cared about open sourcing anything, why are their non-binary Linux drivers so wretched? This looks like more desperation than forward progress.
 
AMD's big push for 2016 is going to be getting a bunch of new "open source" features working for Direct X 11? And if they really cared about open sourcing anything, why are their non-binary Linux drivers so wretched? This looks like more desperation than forward progress.


Well, there is something that will span both Linux and Android.


Vulkan.


I wonder how easily or well this suite of visual effects will scale to mobile solutions?


Samsung exynos chips and qualcomm chips likely sell in the tens of millions if not hundreds of millions of units.

And having a common cross platform cross vendor suite of effects that solves common rendering problems and can be tweaked for specific platforms/hardware seems like it could be compelling.


You can't use part of the gameworks suite to solve and offer an advanced graphical effect on a samsung phone, but if someone creates a game on android that uses a novel rendering technique to get more performance on the limited hardware using the open suite of tools, that knowledge could be added to the whole and ported to other platforms to even greater effect... maybe.

Not sure if that's how it would actually go down, but I hope it will.


I never disliked the IDEA of a gameworks, a common set of rendering tools and suite to act as an easier middle wear for developers. But the locked down nature of it and the blatant use of it to privilege only their own hardware made it a no go for my ultimate dream.

A world where a suite of tools is something the entire world could tap into, alter, tweak, and re purpose for their own needs and others.

Physics calculations is not just an nvidia problem, same with rendering hair or waves or advanced lighting. The idea that there is not some suite that everyone can already tap into and improve upon seems like such a waste.
 
Analogy time.

AMD GPUOpen is to Gameworks as to Gandalf the White is to Saruman.


https://www.youtube.com/watch?v=Dl2R4v9YvEA#t=1m50s

whole clip with the start of the relevant section, but the money quote is right after 2:45


AMD: I am Gameworks, or rather gameworks as it should have been.
AMD will launch their GPUOpen suite in January and then it will fall off the face of the Earth for a year, maybe a few years. We'll see their ShadowFX implemented in the next Tomb Raider game and then, poof, gone again.

The libraries might be good, probably even better than GameWorks, but AMD has no way of encouraging developers to use them. It's even more problematic when developers have to choose between GameWorks & GPUOpen... Make no mistake, with Nvidia involved, there will be a choice. And I don't think many developers will take open-source code they have to adapt to their game themselves rather than allowing Nvidia to do it for them.

You can't have your cake and eat it, too. The more mass marketable a set of libraries like these are, the worse they will look and/or run. Neither Nvidia nor AMD have the money to develop things like smoke, fog, water, and physics effects on a case-by-case basis for every AAA developer. There's a reason these developers don't design the effects themselves -- Because it's expensive, and time consuming.

So what you have here is AMD preying on the unicorns and fairy dust inside their fanbase's imaginations. Either GPUOpen will run very well, and look gorgeous, but only be featured in 1 game per year, or it will be in tons of games and look or run like shit (just like GameWorks). People act like Nvidia is fucking around with GameWorks, if it were possible to make it run better without blowing tons of money or development hours on it, Nvidia would have already done that. It's just not possible. Unless a company exists for the sole purpose of providing these libraries at a cost. Currently, GameWorks is free as far as I know. Nvidia implements it as a marketing tool. GameWorks doesn't run badly because Nvidia is "crippling" hardware, it runs badly because nobody cares enough to make it run better. It's not profitable enough.

AMD got lucky with Mantle in that there was no competition from Nvidia, or anyone else for that matter. DX12 wasn't even announced at the time! It's easy to accomplish something when you literally have no competition in the market. TressFX has been "open source" for a few years and has had no exposure beyond the original Tomb Raider game and an Indie game I forgot the name of. It hasn't worked in the past, it will continue not working.

The announcement today is a marketing move. GPUOpen's existance is also a marketing move. Judging by some of the comments I've been reading today... It's working. Nvidia should announce a new open-source project they have absolutely no intention on finishing and earn some good karma.

I would like to live in the world AMD has painted for us. But if it were a reality, everybody would be going bankrupt and everything would only exist in press releases.
 
Last edited:
Nvidia will use them, and add them to game works!! and then people will cry when they run like crap on AMD cards since NV made some changes to them.

then we'll look back at this thread.

the end
 
There just tools that can be used rather then tools that should be used.
Maybe the PC gaming industry does not see it as an advantage to keep developing these tools because they can not claim copyright on it, which is very shortsighted...
 
Interesting. It is nice they are keeping it open source. But if you ask me they should make it closed source like Nvidia and do they same shit they do. Might be the only way AMD will ever bounce back is to fight dirty like Nvidia.

Either or. Just sounds like a new name for AMD gaming evolve or whatever it was called.
 
Interesting. It is nice they are keeping it open source. But if you ask me they should make it closed source like Nvidia and do they same shit they do. Might be the only way AMD will ever bounce back is to fight dirty like Nvidia.

Either or. Just sounds like a new name for AMD gaming evolve or whatever it was called.


They don't have the marketshare to keep it closed source, it all comes down to development dollars, which they don't have.....

Yeah I don't see this going anywhere unless they are more dedicated to supporting it. No company in their right mind will help make something like Gameworks and then just give it away. I can see individual effects being done that way, (they will probably be unoptimized, just like AMD's examples they have now) but not a full physics and graphics libraries, that's asking for a lot. And this is why none of AMD's initiatives never took off, they were/are expecting something quick and that will never happen, it takes years of dedication, years of solid marketshare, years of support.
 
Interesting. It is nice they are keeping it open source. But if you ask me they should make it closed source like Nvidia and do they same shit they do. Might be the only way AMD will ever bounce back is to fight dirty like Nvidia.

Either or. Just sounds like a new name for AMD gaming evolve or whatever it was called.

AMD can bounce back by launching new cards. They missed an entire release cycle by rebranding existing GPU designs as new products for the various R7 and R9 3xx cards. Fighting dirty won't get them anywhere when people consider the product to be dated and power hungry.

Or they could just get down the price on the Fury cards, I suspect they would sell a lot more with a price drop.
 
I'm not holding my breath. I REALLY want AMD to be competitive again, but I know AMD's (mis)management history: AMD will abandon this as soon as a new executive takes the helm and the whole project will be completely irrelevant in a year.
 
They don't have the marketshare to keep it closed source, it all comes down to development dollars, which they don't have.....

Yeah I don't see this going anywhere unless they are more dedicated to supporting it. No company in their right mind will help make something like Gameworks and then just give it away. I can see individual effects being done that way, (they will probably be unoptimized, just like AMD's examples they have now) but not a full physics and graphics libraries, that's asking for a lot. And this is why none of AMD's initiatives never took off, they were/are expecting something quick and that will never happen, it takes years of dedication, years of solid marketshare, years of support.

I could have sworn that Frostbite 3 is using Mantle and that was not big enough for you somehow?

You have a very skewed look on things, we have seen that with Batman :AK steam is offering refunds until the end of the year even with _all_ the help of the Nvidia developers. This was very well documented.
 
This is awesome!

Gives the devs options instead of binding, locking them down, and muzzling them.
 
I could have sworn that Frostbite 3 is using Mantle and that was not big enough for you somehow?

You have a very skewed look on things, we have seen that with Batman :AK steam is offering refunds until the end of the year even with _all_ the help of the Nvidia developers. This was very well documented.


Look if this turns out to be shit just like all of AMD dev intiatives it won't do anything for them, just like their rewrite for linux drivers (twice), just like their rewrite of OGL drivers (twice) and open sourcing them both Linux and Windows OGL drivers did nothing for them, just like Game Evolved, just like AMD get into the game program, just like their GPGPU open source program, it doesn't matter if its open sourced or not, they need to SUPPORT it. $$$$$$$ is what makes support happen, not mental masturbation at a fanciful idea to open source, as if it would be a cure all to actual financial spending in supporting a program.

The pathetic thing is these are just slides! Will slides do anything for them? Oh yeah a dev needs presentation slides of something in the future? But they are working on their programs NOW, what does it do for them, nothing, they aren't going to hold their dicks in their hand and wait for AMD to get moving. Oh yeah how many years has it been since Gameworks has been around, and how many years of tress fx? How many years has TWIMTBP been around, how many years has AMD been working on GTG? Did tress fx and GTG do anything compared to the Gameworks or TWIMTBP?

How many slide presentations have you seen nV do for Gameworks without an actual alpha or beta to show for it? I can't think of one time they did that. They would at least show a video of what they are doing, it might not be as "pretty uped" as a full game demo video but they show it, and dev's have access to it if they want to use that particular library at the time of announcement. Not a year or months away.

Thank god people like you don't do anything that remotely connects to the game industry because it would die if you had your hands on it just like it did in late 80's.
 
Last edited:
if anything comes out of it it will be more then just a shader library if done right it will be a fully functional 3rd party program that can be easily be integrated into an existing product.

So this is what it needs:

1) Shader libraries
2) wrappers and classes that are easily accessible
3) ease of accessibility should not prevent or reduce already done optimizations within the libraries
4) documentation how to use wrappers and classes to the full potential of what they are ment to do.
5) scripting language if necessary, probably not since its open source.
6) Integration with current top engines on the market
7) full set of testing tools which support the shader libraries and that can again be integrated into an existing product and they should also be stand alone tools for test analysis.
8) documentation for this as well
 
if anything comes out of it it will be more then just a shader library if done right it will be a fully functional 3rd party program that can be easily be integrated into an existing product.

So this is what it needs:

1) Shader libraries
2) wrappers and classes that are easily accessible
3) ease of accessibility should not prevent or reduce already done optimizations within the libraries
4) documentation how to use wrappers and classes to the full potential of what they are ment to do.
5) scripting language if necessary, probably not since its open source.
6) Integration with current top engines on the market
7) full set of testing tools which support the shader libraries and that can again be integrated into an existing product and they should also be stand alone tools for test analysis.
8) documentation for this as well

I think the success of this depends on how much initial "tooling" AMD puts into place (and the quality thereof), how much support AMD is willing and able to provide, and if a dedicated, self-maintaining community ultimately grows around this.

But looking at historical evidence about AMD's "innovative" ventures, this is just another marketing publicity stunt...
 
I think the success of this depends on how much initial "tooling" AMD puts into place (and the quality thereof), how much support AMD is willing and able to provide, and if a dedicated, self-maintaining community ultimately grows around this.

But looking at historical evidence about AMD's "innovative" ventures, this is just another marketing publicity stunt...



I agree, and it will take time, something that will take a few years its not a one off shot that will do it for them.
 
So when can I expect to buy games with GPUOpen? And why is AMD using the term "GPU", something coined by NVIDIA? Wouldn't it be more proper to call it "VPUOpen"?
 
It's just another option for a developer to use. I continue to dislike all these competing standards, it only serves to fragment the pc gamer base more.
 
So when can I expect to buy games with GPUOpen? And why is AMD using the term "GPU", something coined by NVIDIA? Wouldn't it be more proper to call it "VPUOpen"?

Well according to the article and the quote in the OP it launches in January so some time after that I'd guess. Assuming that wasn't just a criticism disingenuously dressed up as a question mind you.

Regarding the GPU "question" I'd suppose it's because it's the more widely understood term. Granted Nvidia did popularize the term GPU but it's just an acronym not a Polio vaccine so really, who gives a shit?
 
Well according to the article and the quote in the OP it launches in January so some time after that I'd guess. Assuming that wasn't just a criticism disingenuously dressed up as a question mind you.

Regarding the GPU "question" I'd suppose it's because it's the more widely understood term. Granted Nvidia did popularize the term GPU but it's just an acronym not a Polio vaccine so really, who gives a shit?

The question was legitimate but it is funny that AMD tends to always follow where NVIDIA leads. FreeSync and now GPUOpen, they're like those generic cereals you see in the aisle of WalMart with terrible packaging and mascots that are rip offs of the real thing.
 
So what exactly is this?
An open-source library of shader code?

No, it's AMD's PR slide announcement of that they would really, really like to have an open-source library of shader code.

Now if we all pitch in AMD would be very happy.
 
AMD looks always great on slides, they almost got me with the Fury X ones, the only thing that held me back was the coil whine issues from the card and pump.

Wondering if this will end up like a Bullet&#8482; drop
 
Does the use of GPUOpen come with the same support that Nvidia provides for Gameworks or will they cheese back to the MIT license disclaimer?
 
The question was legitimate but it is funny that AMD tends to always follow where NVIDIA leads. FreeSync and now GPUOpen, they're like those generic cereals you see in the aisle of WalMart with terrible packaging and mascots that are rip offs of the real thing.

Yet, nVidia follows AMD on the important stuff, like GDDR5 and HBM.
 
Back
Top