KWin May Drop Support For Catalyst, Vintage GPUs

[aceman]

So there's no point in continuing to support legacy Catalyst in KWIN, unless I'm missing something..

I'm assuming by legacy, he means the hardware which is no longer supported in the latest versions of Catalyst.

According to the article it would drop support for ALL Catalyst users, even with new hardware (like HD7970), because they use it as OpenGL 1.x for speed reasons. See the quotes in the article.

[FireBurn]

You must be living under a very large rock...There's been dozens of Mesa 8 articles...

And how many tests have been run that actually use the new GL3 code?

[hal2k1]

And how many tests have been run that actually use the new GL3 code?

Since the entirity of the Phoronix test suite worked with code that ran under Mesa versions prior to Mesa 8, it is clear that none of the software uses the new OpenGL 3 capabilities of Mesa.

However, it is entirely possible that some of the software (especially games) could detect that it was running under non-mesa stack (fglrx), and hence automatically switched (when using the proprietary driver) to utilising more advanced OpenGL features.

It may well be that the Phoronix suite in some cases merely reports on the difference between OpenGL 1.4 versus OpenGL 3 or 4, instead of reporting on the differences between Mesa and fglrx.

[cl333r]

Yes! Remove old crap!

Stop supporting it and some people will realize they live in the past century.

[e8hffff]

1. Open a ticket on the public bugtracker. I think we all know how much good that would do.

And this is the problem with Open Source concepts. There needs to be assigned project deligates, who promote and request movement for the project teams, to say AMD, or to promote more open coders to come in.

[GreatEmerald]

Even though I'm a fglrx user, I can't say I care about that a whole lot, since for me fglrx has been corrupting the screen when KWin compositing is on anyway (black screen areas due to their new Direct2D implementaion)...

[smitty3268]

And this is the problem with Open Source concepts. There needs to be assigned project deligates, who promote and request movement for the project teams, to say AMD, or to promote more open coders to come in.

I'm not sure what you think Open Source concepts has to do with fglrx. It's about as closed a project as there is.

[asdx]

This is a great decision and I support it. Thanks for moving things forward.

[smitty3268]

On the blog:

the current code will stay at least for another year

I'm not 100% sure what that means - it's probably just referring to the January 2013 release, since the one for this summer is already branched.

[slojam]

This is a bad decision and I don't support it. Thanks for moving backwards.

Drop fglrx, and I will drop KDE. hmm yes maybe its time to move on.