Typical journalist from "pro" site, for You.
APIs aint copyrightable (whatever any non-programmer could tell You), so they can not be proprietary nor "open sourced".
Open as in Open Standard or closed as in developed by single entity.
So its "open API" which for now means "PR" or "open source TOOLS/LIBS", which would be nice. There is also chance that Mantle code is different from Catalyst, and AMD really want to make it open source down to the last driver bit. They already have MESA, so they have that skill set, and they have many times told us that they make each GPU gen more "open source" safe, from their "company secrets" point view. And that would play nice with AMD push for CPU like openess about what happen on GPU.
Talk is cheap. Release something.
If this is as good as AMD's proprietary drivers I hope it never sees the light of day.
I am uber sceptical on this...
I'd love to call myself AMD GPU fan, but it won't happend ny time soon. I am using their processors, for as long as I remember, I only had intel processor once on my personal computers.(2nd gen celeron)
First of all, AMD drivers suck elephant balls, probably blue whale balls. Tests on Phoronix prove it over and over again. A stronger by means of hardware and more expensive card from AMD is worse than one from Nvidia.
And at this stage, not even promise of open source drivers will attract me to AMD.
Finally devs agreed more or less to use an open standard like OpenGL and not DirectX and boom, we got also Mantle. A half backed standard, that will most likely NEVER will be supported by NVIDIA. So it's like 65% or more of Linux world out in to the window. And it gets worse from here. Knowing quality/performance of binary and OSS drivers from AMD, I DO NOT BELIEVE mantle is gonna give any advantage, performance wise, to any dev on Linux. And even if it will, it will probably be available to one or the other type of driver + not all hardware will support it. Even if I am partially wrong here, how many people on Linux will be able to take advantage of it? 5% from already small Linux user pool?
I honestly hope that AMD will fail with Mantle, quick and painlessly. So they put more effort into existing OSS drivers. They need to do OSS drivers first. We were fed with promises for ever.
Saying, that OSS drivers "work", IMHO, is the same as saying, "I bought 50k Euro expensive car, it only goes 50 mph". IMHO, this car is broken or too bad. It's not worth 50k Euro. Pick and choose, either broken or overcharged and outdated.
AMD, get your stuff together please. I never had AMD card in my life, but I want to have it!
Mantle aint either-or with Mesa.
Originally Posted by dimko
Though I hope that gallium is able (will be modified to) support such API.
That is the future of computing on GPU (engineering decisions behind API design), so lets prepare for it.
Call me when Intel and NVidia support it, otherwise this will only mean more complications.
The U.S. Federal Circuit Court disagrees with you there.
Originally Posted by przemoli
For what I read after mantle presentation, some Features where needed which on AMD hw were introduced with GCN, but on other vendors GPU the features should be present on much older hw.
Originally Posted by Nille
Sadly Mantle will not bring any games to Linux. Who wants to cater to 4% of 2% of gamers, where only 10% of those would actually buy the game?
And no. Mantle doesn't simply need DX11 hardware. Even the DICE Mantle guy said that he doesn't know if it would work on Kepler.
They won't. Consider the differences between OpenGL and Mantle. If for some reason Intel and Nvidia support Mantle, then future versions of Mantle will rely on AMD, like Mantle 2.0 or 3.0. That's giving a lot of power to AMD to say what new features should be included in future versions of Mantle. This is why it's better to have an API designed by Khronos or Microsoft, even though Microsoft does put Xbox One over DirectX for Windows.
Originally Posted by volca
Mantle uses HLSL, which seems to share something with Direct3D. Maybe beneficial for Wine performance on Linux?