AMD Is Still Moving Towards A Unified Open-Source Driver

[dungeon]

But you are now forced to use a bleeding edge kernel for newer cards.

Mostly yes if it will be just like current radeon, but we don't know yet for sure how unified driver will behave... it is not there yet

[Nille]

there's absolutely zero reason to keep catalyst closed source when mesa is catching up so fast, at this point it has to just be a duplication of effort

Linux and OpenGL is obvious not AMDs Main Target.

[Nille]

Mesa's catching up because AMD Developers are coding for it.

Can you Link some of the big changes in mesa from AMD? I read mostly only that Intel has implement feature xyz and the other free drivers catch up this features.

[Nille]

Which is great, Catalyst users will never complain again that there is no kernel support in time nor need to compile module

But you are now forced to use a bleeding edge kernel for newer cards.

[dungeon]

For the record - if you really had to do that, likely fglrx wouldn't be good enough either. There's a reason everyone develops GL apps on nvidia first.

But this one is interesting to me so i must , firstly you say OSS driver is enough for me now you say use nvidia first . What you have against fglrx?

"Do not answer i already know" what kind of bullshit is that

[bibaheu]

This might bring CrossFire to Mesa... Imagine "CrossFire" with totally different GPUs using Mesa, it would totally rock.

[dungeon]

I am going to assume you aren't actually a game developer. Nobody at this point would develop any game or engine which has 4.4 as the minimum version. Every real game or engine being developed WILL have fallbacks to support older versions.

No i am not (and i already mentioned that ) a game developer, just trying to realise why fglrx is needed.

You are spouting nonsense. I realize you don't see that, but you are. Which GL 4.4 feature do you require for your engine? That can't be easily turned off?

Or are you just trying to spout some theoretical argument that has no bearing in reality? Don't answer that, i already know.

For the record - if you really had to do that, likely fglrx wouldn't be good enough either. There's a reason everyone develops GL apps on nvidia first.

We are talking about (and i only make XYZ situation when fglrx is needed) why should or not AMD needs to abandon fglrx for Linux . And only stay or not with only opensource driver

[smitty3268]

In short, it sounds to me like this comes down to the definition of "engine".

If by that you mean, a toy application you are personally playing around with as a hobby, then yes, your arguments start to make sense.

If you mean a real, actual engine you expect other people to use in the real world, then no, you aren't making any sense.

[smitty3268]

4.4+ so OpenGL 4.4 will be minimal version How can i work with those things today if i don't have fglrx?

Again i am not developing any engine, but if i start now to develop that New Age engine fglrx will be the only choice

You are spouting nonsense. I realize you don't see that, but you are. Which GL 4.4 feature do you require for your engine? That can't be easily turned off?

Or are you just trying to spout some theoretical argument that has no bearing in reality? Don't answer that, i already know.

For the record - if you really had to do that, likely fglrx wouldn't be good enough either. There's a reason everyone develops GL apps on nvidia first.

[RahulSundaram]

4.4+ so OpenGL 4.4 will be minimal version How can i work with those things today if i don't have fglrx?

Again i am not developing any engine, but if i start now to develop that New Age engine fglrx will be the only choice

I am going to assume you aren't actually a game developer. Nobody at this point would develop any game or engine which has 4.4 as the minimum version. Every real game or engine being developed WILL have fallbacks to support older versions.